Tech Innovation & Social Impact
Hi. Welcome to another Wai involve podcast. Joining us today are Parag and Kirti Bihade who are leaders in technology and co-founder’s of Datasmith AI and Parallel Minds, Parag brings in over 20 years of expertise in digital transformation, custom software development, data engineering and AI integration. His experience spans a variety of domains such as manufacturing, oil and gas, healthcare and e-learning. Kirti on the other hand, is an accomplished and seasoned data science professional with over 15 years of experience. Her expertise lies in generative AI, machine learning, and computer vision.
She is also the founder of Swapnapurti Social Foundation, which is a platform that is dedicated to empowering rural jobseekers and driving social impact. So they are in a different league, but together they blend cutting edge tech with traditional methods. Their solutions are impactful and empower communities. I'm so excited to hear more from them.
Welcome, Parag and Kirti. Thank you for being here and welcome to our Wai involve podcast, Parag , Data Summit and, Parallel Minds. They are known for blending technology with AI integration and developing so very innovative applications. So, would you like to walk us through, a project, a stand out project, which you did, where you got significant and amazing results? Yeah. I mean, we can certainly. By the way, thanks for inviting us here.
It's our pleasure to be . kirti and I I am so thankful so much for you guys. Thank you very much. To wai It's a warm feeling, talking to you guys here.
Great initiative though. So. Yes. So coming back to your question. So we have been, doing interesting projects, in the real estate space or in the oil and gas space or typically in the health tech space and so on. But I will pick up one particular project, where, it's in the health tech space.
In that space, what we have done is there is a customer in the US, he runs his own hospital, but almost 80, 90 people he's working with, all the doctors are working with him. And, one of the problems, he realized is a nephrologist. And one of the problems he realized and identified in the US that the there are a lot of patients, senior patients, I would say elderly people who have dialysis issues. Plus, the problem with in the US, it looks like or in typical outside of India scenario, looks like or even in India these days? What it looks like is people have to queue up to get to their, particular, slot where they can get the dialysis. But then dialysis is not time dependent. When you need dialysis you need dialysis.
You need it. Right. So there is that is where then he really was looking at options. How do we really cause this and solve this problem. And that is where he came up with this idea of, remote, dialysis, where, the idea is such that that you will have a home dialysis system at home dialysis system at home.
And, but then as a patient, you got to monitor you got to monitor the vitals right. And you cannot on your self. You cannot simply straight away get into the dialysis mode.
Right. So that is where we collaborated with him. And then we created a Ideated a solution. The solution was based on the smart glasses. So the smart glass, is an interface using which you can as a nurse, for example, who can visit at home. The nurse through the glass can look at the patient, and then the nurse can give a call to the doctor who is in the remote mode.
And on the web mode. The doctor can easily interface, look at what is happening with the patient, and then through voice, he can guide the particular nurse using the glass. And then then he can direct. The doctor can direct to the, to the nurse how to operate and how to make things work for and all that.
Right. Fantastic. So it's all a combination of web application, combination of, transfer and smart glasses, combination of, AR enablement and so on. So all that is all that is something which what we did for the customer, that is one of standout projects. It's amazing. Yeah, yeah, it is solving a major problem of elderly care. And and then the nurse, you know, using the technology and then doctor in the remote mode, really guiding the nurse to, you know, make the operations happen there.
That is something, solving a problem. So what kind of generative AI, related implementation did it require? Not in that project, though. But then, generative AI for that, maybe, you know, we will have to get some inputs from Kirti.
Yeah. definitely on the Generative AI kind of use cases which we have built and what we are working on it has been last couple of years, it hasn't been a total game changer. So, Datasmith AI is other company people are like founder and co-founder to that company.
So it was completely the theme was that this company will work with data and we apply machine learning, deep learning, data analytics, visualization here in this company. So earlier when we started so Datasmith AI started from 2021 and Parallel Minds. since 2008. So when we started this company at that time Gen AI was not so popular. So we started with that vision that we will apply ML and all that. So then we have done very interesting projects.
And there were customers from Parallel Minds from oil and gas domain. And they were having good experience of their domain and fun machine learning also. Then we got very good projects in the oil and gas domain. So we work there in the heavy data analytics space and machine learning space, data visualization space. So we started there. It was going good analytics and machine learning.
Then after that GenAI came into existence, and then we saw that GenAI is really picking up very fast. It has really good business scope. If you are considering of, company from a company perspective and even we were receiving so many use cases from the existing customer from Parallel Minds. for example there are like real estate customers, they were getting multiple brochures, multiple information and tenant information, and they were extracting that information manually. So we got use case from that side.
Then there was another customer, for example, compliance related stuff and all that. So there are important information in the compliance that if you will not follow this compliance, there could be a possible loss in the future. So that kind of important use cases we started receiving and then we thought that this is really a good space to go in of course we have that, competency of ML, but it is good space to one. Then we started Generative AI. And now when we thought that we will position Datasmith AI as a Generative AI company, then again we came across one use case which was really interesting.
And from this use case we have built our product and it is really doing well. So do you want me to explain that product? maybe later I can explain that product. You can you can do this probably we can. I mean just give a brief overview of what the problem statement is, what is identified and how we have been able to solve it.
So typically, imagine a valve manufacturing company and they do receive tenders or RFP, they call it request for proposals from oil and gas companies let say Aramco or Adnoc or Sharjah National Oil Company and so on. So once they receive tenders, what they have to do is they have to go through the tender. And tender typically is a document of 400 pages or 800 pages or 1200 pages or even 2000 pages. Now, because the sales team is a cost function many times and it is, and there is a limited number of people who are in the sales team now, the sales team members, let's imagine there are five sales team members. So they have to they are receiving this tenders on and off all the time. So every day they are receiving, let's say, 8 to 10 tenders and imagine the amount of load or the time they have to spend going to the tender and identifying whether this tender, am I supposed to respond to this state? Yeah.
If the answer is yes, then there is a timeline to be met. And if the timeline is supposed to be met, then they have to refer to a lot of things from the past experience. They also have to go to the tender and and analyze it properly. Try to find answers to the most important questions, identify ambiguous statements.
And so many regulatory clauses. ASMA standards so on. And then eventually they have to arrive at what content to be written for each section in the given proposal.
That is what their challenge is. Now, the problem is because it is a limited set of members, they just lose on opportunities. So there is a huge opportunity loss they are at, and they leave lot of money on the table. And that is. And the other challenge what we have discussed is there are juniors, in the team who are not able to concentrate and they do not even understand how to assess a particular RFP or a tender because they are juniors.
They don't have that acumen or the understanding of how to read the clauses. And that is where there is a bottleneck, where the senior person is just a couple of people and there are so many juniors, they are not able to concentrate because there is so much regulatory stuff. Yes, it's easy to, you know, not detect anamoly anything that is missing, right? So that is where the tender genie comes into game. So what they said actually their first basic pain point that sometimes there are so many tenders that there is no time to look into tender and check like last submission date.
Yeah, even there is also only based on that. Also they are losing out on their thing and even they ask. For example there is some keywords like UL is a keyword.
Having this keyword can cause them like if they are missing out on this keyword, they can. There is a loss of one crores 2 crores, so it is important for them to find out all such keywords, all such important information. from that So we built this system to analyze tenders, and you get answers to the same set of questions. which they might want to get answers to from every tender they want to analyze, to analyze with, help them create proposals, help them create winning proposals is what our goal is great. And, it's a great solution. They also want to they so they liked the solution so much and they said was instead of, one tender at a time, why don't you automate this process? So we created an automation solution where they don't even need to come to our conversational interface anymore.
They just send the tender documents to one of systems, imagine one of interfaces, and automatically all processing happens in the background magically. That is why we call it Genie. And then the genie does it's magic. Applies all the logic. Have AI ML we have Gen AI are all combined technologies together. And then we generate a document, generate a proposal to them, and then we send out back to their set of team members.
And the entire process only takes like 30 minutes maybe. So 1800 pages. A document is done in 30 minutes. And it's not just about reading the document.
It's about analyzing, assessing, coming up with reference documents and all that. So your experience and expertise lies across so many different domains like healthcare. You talked about real estate oil and gas And I'm sure every industrial domain has its own challenges, certainly in terms of AI implementation. So in your experience and in your perspective, which is the industry that that has most, challenges regarding AI implementation and adoption, what do you think? I think I will take that question. So generally, what is happening the AI people are like they all started believing in it earlier.
There was some resistance that it was just hype and nothing is happening on that. Right. And then suddenly ChatGPT when started, people have started seeing the results and they felt that now it is doing something right and then eventually text data, audio, data, video, data. That is one good thing. Everybody was started at least believing now when they started applying that use cases, right? So whenever a genie is providing solution, it is a value based.
It is an efficiency which directly you cannot say to the customer it is going to add revenue, because generally what is happening with any business, if there is a cost going only and not coming anything, it is very difficult to convince customer that this is going to work for you. Absolutely. So they know AI they are believing in AI, but whenever it is coming to cater to their use case, they want to develop POC, but they are not going beyond POC . A certain disbelief in AI related decisions because it has to be human in the loop. Yes, always the final decision and but interesting. Sorry to interrupt. Yeah. Go ahead.
That's that's okay. I mean, you can go to the interesting scenario on this last particular point. I would like to share a quick story on how we picked up the customer.
so one story is so we tied up with the with our partner in Australia. Real insight is the name of the partner. And together, we pitched a customer in Gulf country.
They had almost, 50 wells oil wells I think, like 22 oil wells and 137 latterals. That is a subwell we can say that. So, so in a typical oil oilfield. Right. So you have wells. The wells could have different profiles and all that.
And then wells are additionally directionally drilled and they become horizontal, wells and so on. So they had this data of 20 last 20 years. So be data. So be it seismic data, be geo data, be it lithology information, be drilling information, be it so many different file formats and so on.
Now the customer what they wanted to do is they have their own geologist team, they have their own engineering team. And they wanted to take a call take a decision. That so let's say there is an oil field and there is another oil field just besides the existing oil field with similar characteristics of the, the that is, we they call it as a bio static bio stratigraphy. So imagine that. And what they wanted to do is they wanted to take a call whether. So analyze this data and probably figure out where I should make the drilling decision, whether I should drill here, whether I should drill there or there were.
Imagine some 4 or 5 wells already drilled and they wanted to take a call. In what direction? In what direction they should drill. Okay. So now for that, they had done all their job and they have taken the decision that these are the points where they want to drill.
But someone from the customer side probably thought, let us apply machine learning. Let us figure out what can I predict? The same results the way our engineers will and that is where we saw this positive, I would say opportunity where customer of that scale oil and gas domain, they are wanting to experiment and figure out and validate whether machine learning can also do the similar prediction as what the humans have done. And was it near the human accuracy was 86 percent They were really fantastic.
And that is the project where we realized like, you know, fire. And so so our model was able to predict it is 86% is the rate is the is the is the accuracy of the prediction. And then they were able to compare they had actually presented this case in ADIPEC there is a conference Abu Dhabi International Petroleum Exhibition and Conference And they also written a paper. Published a paper. She was also part of the the content writing team and all that. And eventually when we presented this particular overall story to set of additional customers, they made us realize what kind of value we have added in the customers.
And so that is, we are $5,580 million of improvement. Is value is what we equivalent value is what we have achieved because of this? Yeah. Enablement of the process. So so humans and AI can absolutely work together. It's good for both. always.
It is it is not only I generally what I mean, you cannot say, just AI with the final decision there is always the human element. It's going to be there. It is actually final stage where you are sort of validating the decisions that AI has taken. So that was one of case. So yeah.
So always AI what, what I feel like AI is not to generally people say will replace people or there are like manual work people are doing for example. And they are taking 32 hours and by oh doing that work in one hour we will replace their job. It's not it's not like that. It is to empower them so that they can really work on important, important aspect that how they can win the proposal that is one aspect.
And what is another aspect on this particular project that then pain point is this new comers that is new joinees. or young people you know, their attention span is very low and this monotonous task require a lot, lot of attention. So details are very important.
So what they are saying these people are not doing their job and they are not interested in doing this monotonous job. So eventually in the future they are going to face the problem. Yeah. Right. So this is where we should help people.
This is where AI should come. The data is showing the right results. Right. If if it is just the you just the intelligence probably people would doubt that.
But it is your data. It is your line of business data. You are looking at and patterns are emerging out of it. So you are seeing what you are actually working. It adds tremendous value.
They are now feeling so confident that they won't lose out on RFP. They would look and look into such RFP, which are having higher winning rate for them, that the second thing they are really confident on. Plus we have been able to because we are a boutique, yet we are still in the making process, process, in making the product, in making so what? Another thing which I would like to highlight is she is of very strong opinion that we should not get trapped in one more feature that typically our product, we typically, as a product owner, you challenge to make it will be cutting it. Could you sub select our subcategory.
And that is what we see. So wish list of features that never ends that you are operating from my previous earlier experience of deals and offers.com we said nothing doing. She also took a stand. Nothing doing. Let us go to the customer first. Align with the customers. customer has to give you those features.
Otherwise it is going to be a monster of an application that so many features and we do not know whether, all the features will be utilized by the customer or not. It is like feelgood factor for us. Yes. Yeah, but this wishlist never ends were it it never ends. So many applications are floating around and we are always attracted to new feedback. Right. So that does help.
The customer is happy. it is important Now what. Because generally we are getting so many things. AI is Generative AI there are so many ready libraries. are into the picture so building solution is not a challenge right now right any solution. But what solution you should build and which use case you should cater that is more important.
One also be customer centric. Absolutely you are. You are solving a problem that the customers face. What we are trying to convey to the customers is that we are not a POC company can be our our bold statement on the website is can we go beyond policies? And with that, we also approached, you know, enterprise customers, where we presented our case and they really liked. And then we have this program called us Paratroopers program, where this young talent we because she was a professor for so many years, she had good access to a lot of engineering colleges and as well as MCA Colleges.
And she was because she has done in AI. So what has happened is she really understands what skill set or what kind of topic awareness a student should have. So we are able to identify some colleges, not only in Maharashtra but across India where there is specific course is being taught. So we hire people from those particular colleges. We groom them, train them, we invite them, we really groom them, and we make them productive in hardly 4 to 6 months time.
And those are the people who have been able to contribute in the developing of our product. So that idea is what itself, the customer like of that skill. Like, like there are juniors, and there are seniors like us. We have been able to groom these people and build the IP, which is adding value to the end customers.
So so that customers said, we know why you want to be a hundred people team for us. Not too bad. Let me ask you this. Can computer science graduates fresh out of college directly get into AI? What skills do they need as a foundation or what? What is the grooming period for getting into like an AI engineer? I think kirti will be the right person to answer. Yeah. So not just this.
If there is a computer science also there are few subjects, added into the syllabus at least basic machine learning concept, deep learning concepts. So when you go in any college these are the subjects in computer science field also okay. So depends on the students few. Students are very curious and they expand themselves other than this stuff.
But if you consider that only computer science student need to know background of machine learning or deep learning, I think it will take almost six months of rigorous, work to arrive at a good AI engineer, at least to start with. His job in machine learning project based work or just a learning three months will be learning and three months will be a project based thing, then only we can learn in a good way and he or she can enter into the industry. So that is what I feel, because learning AI means you need to start, really start from scratch. And statistical analysis, linear Algebra, then machine learning.
Then you should know the NLP. Also, if you want to get into generating AI and then now there is a GenAI and again, there are different, lines into this area like till we can groom that particular candidate that whether he want to enter into this only machine learning field or you want to enter into the generative space in that way also we can direct, the people and then they can go into the respective fields. So kirti you have been in data side on the data, data side for so long. You have expertise in data engineering, you know, so you must have, connected so many different line of data, in complex domains do you want to walk us through any project that.
Yeah. So I will walk into the projects with good insights and that's, so generally if you go on LinkedIn. So I have that line that, I am very passionate to bridge the gap between domain and data. So what is happening for example, oil and gas. We will take example of that. So those people are very rich in their domain because they know in and out of their domain.
And we are rich in data science, for example, that we know data science, but who will help them because they need one person who will understand their terminologies in their domain and knows this data science field as well. Then only we can combine the power of two fields and we can come up with solutions. Before I have started my own company, I was working with one other company. Now it is acquired by Ciklum Company so there I have worked with data where the there was a data from kids like of the age of three to for example seven. It was related to the like they were playing games. In terms of educational games, we can say that they were like numbers one, two, three, four.
And there were questions and at the end of questions they have to pick it like, right, wrong, whatever. And we were getting analysis of their data that 3 to 4 years. What is their problem. Are they finding any particular letter difficult. Are they finding any particular sum difficulty in the maths so that eventually we will give this insights to the game team so that you can just, this question is more complex because maybe 90% students are not able to solve this problem.
And all that, but eventually in that insights, what what outcome, we observe that the number one four is, is very difficult for any kid, for example, maybe 3 to 4, five years. They are finding it very difficult to identify that number four. So we thought why why it is like 12398. They are identifying variable.
But why number four. Like what is complex in that number four. Then then we, we were reading one research paper into that because in number four the diagonal line there is horizontal line, there is vertical line. And when we are in the nature, nature is surrounded by most of the horizontal line were diagonal is coming later, for example. So our eyesight is more comfortable with horizontal and vertical kind of thing not diagonal.
And for kid to manage everything like diagonal, horizontal and vertical, it becomes difficult for that. And that was then we thought that the research paper is talking about this, that why for us, as even one becomes initially difficult to manage all the angles. And that is why the kids were facing that problem. And later what game team done that they have just converted that into vertical, then horizontal and vertical.
So that kind of thing. Never thought about it literally. It was really very good for us. And with that small change of instead of diagonal line, and instead of making the line vertical, I think, after that, the performance of identifying the, the number four from the same set of kids and even the future kids, they were able to figure out that this is four.
So the simple change from a diagonal to vertical was a very good satisfaction. Then we that is a power like we could identify that why they are not able purely because of data. And the results and the kids were struggling and why, why they were struggling and some making changes in the UX as well as making changes and applying all the science, because that is the point. I like simple thing. I want to talk about the Swapnapurti Foundation, which you have founded. And so when you are using technology driven approach for solving issues, social issues, what what was your experience and how do you do that address the purpose of starting swapnapurti was, I'm from very small village, right.
So I know all the challenges for girls like eductaion if they want to take higher education or for example, girls want to opt for MPSC or UPSC kind of services administrative services, we can say that there is some resistance from parents sides only right. So that is where because sometimes there are very aspiring girls But they don't get out of the village because of all these restrictions and all that. That is where we thought that we should start swapnapurti social foundation, so that we will help such kind of girls to get out of the village. we will sponsor them for their classes, for their food, for their living thing so that in six months they can acquire good knowledge they can go back, they can study, they can do their work. This is what the purpose was. Yes.
Right now, in terms of technology, it is like we are going to the institutes, we are getting, candidates and we are helping them. We haven't added a technology angle into the swapnapurti social foundation foundation right now Not yet. Yeah. Not yet not yet
Why is that? We have I think we haven't reached to that stage today. So swapnapurti recently we have started 2 years back because as she mentioned she comes from a very remote, very small I think hardly 300, 400 families stay in that village. And in her early days, when she was in the school, there was no school in the village. And she had to go out and stay somewhere else and so on. And then she struggled, with monetary issues, and she struggled with during her engineering days.
And then, then then she somehow overcome all of that, and then she came to a decent place, and then she improved and then reached to, the, university, for that matter, and then finished about to finish her PhD, then get into this entrepreneurship mode and so on. So that kind of a journey. Then she realized, we realized was this seems to be a good inspiration for a lot of kids, from that place to getting to basic education and helping them, manage their aspirations and not just get down like that and even initially see sometimes what happens so there is a rayat prabodini in ABC there they are all giving these classes and most of the students are opting for free classes.
So we went there. We asked them, do you have like girls students those who are aspiring but they are not coming to your institution. They filled up your form they are not turning up and all that. So we took all that information from them.
But there was some, not, we cannot say, resistance from their side, but there was something like, well, they were not coming to us to give us that list and all that, but we thought, we have started swapnapurti foundation, so we should do something, at least through the foundation, if you are not getting girls right now, for example. So we have done something for the flood area. We have done something for the individuals.
student I have identified like this is the student i should help him through swapnapurti So that kind of stuff we are doing right now. So we haven't reached maturity level, maturity level. But really, if you ask me, I'm really more, more than willing to help girl, students, those who aspire to go into the MPSC and UPSC, those who are listening to this podcast, I will ask them, you are such kind of girls? So you should, you should come to the swapnapurti social foundation parag or me, and you can reach out to us. We are more than willing to help them. That's wonderful.
Yeah. yes and the share impact that you are creating that is really so heartwarming to hear and appropriate to your foundation. Yeah. Thank you. Thank you. I want to delve into another topic related to AI regulations. So you are working with healthcare, oil and gas.
So all that data that you are handling, what kind of how do you address the AI regulation, privacy and the, compliance regulations related to data for different, customers for different domains? Okay. So this is also how do you handle how does GenAI, handle that response, this really good question and this is important questions, I think, because every time when we go to the customer, first question is always there. What you are going to do with this data is this data you are showing to another customer because there are competitors, right? So that is first question they ask us now when we say data privacy in healthcare, these people manage like HIPAA compliance and all that. If it is coming from the US customer. But for us, for example, if we are working with tender genie right now and they have their own tenders and you are asking them, give us your past data classified. So that is very important for them.
That is their asset right Then if they are giving us that asset, they will make sure that we will keep it very privately. It should not go outside. So of course we are our own, cloud.
Servers and all that. So we are keeping their data in that privacy network only it is not going out of the network, not even. And sorry to interrupt and more importantly, to address your particular question. So even if the product or other solutions which we are building, they are in the SAS mode, there is logical, completely, I would say isolation of the information, even if you are into building an SLM or let's say you are building an LLM, for that matter, to address or create a knowledge store specific to a particular customer with subscribing to our platform.
So that is a logically separate LLM or logically separate vector database, they call it or logically separate set of API calls. And so all that has been very well orchestrated. So in our solution we have always I'm not just Https. These are all basics that they can do direct the based on the subscriber.
Can we redirect this to a particular, instance of based on the customer who is creating and so on. And even within that there are, you know, access rights and there are permissions and there are privileges based on the current context of the user and so on. So that is how that is well orchestrated and well managed. That is how we are being able to achieve all those, and address those data privacy issues. Yeah. And again,
ultimately there is a trust always between you have to build that trust between customer and your side So generally we say like AI Ethics this should become part of your culture. Yeah. And you should able to say that AI ethics is our culture. So then only it will work what that is what we assume. So you have an end to end service model at Parallel Minds for digital transformation. Yeah.
So what sets it apart, in your opinion from other consultancies. Well, fantastic question. How do you maintain the edge because this market is so competitive, because you use digital transformation that has been the prevalent and it has been the established term for so many years. What we are bringing a different sort of changing the mindset is AI driven digital transformation. That's the first change we have brought in the way we are discussing with the customers. What we have observed is many customers, everybody seems to have a AI budget, right? And there are big players who handle big customers, but there is a specific need to have identified in the market.
There are imagine there are manufacturing companies, there are equipment manufacturers, there are typical manufacturers. These mid-size customers who are, let's say, $100 million to one, $150 million to $200 million to, let's say, $300 million. That's in the customers.
What is happening in last five, ten, 15 years is they themselves are getting acquired, or they are also in the mode of acquiring customers, acquiring similar businesses, By worthy of acquisition. What is happening is naturally there are different line of business applications. At each of these acquired companies or plants and they have different CRM.
And they might have different ERPs they might have their homegrown systems, they might have different attendance management system, they might have their different formats of documents and so on. So what we feel is the buying company, the the I would say the holding company is always in the pursuit of developing solution, which will give them a unified view of what's happening in their overall business. And that is where in bigger companies may not come. Service providers may not come to these mid segment market customers. That is where we would come in because we understand AI we understand typical GenAI use cases. We understand custom software development.
We underst, we have been doing workflow automation. We have been doing digital transformation. Typically we do not call digital transformation as transformation. What we do is we work very closely with the customers.
We conduct discovery sessions, we conduct user experience sessions. We find we do conduct gap analysis and then we make them realize that there seems to be a process chaos. There seems to be a data chaos. There seems to be a complete, interchange data exchange chaos. So how about then identifying, a scenario, where then we take that as a case, and then we present to them this could this is how we could be doing it better, or this is how we could be integrating various system. This is how we could be applying technology.
This is so we play more of a consulting role, to begin with rather than today. So we do purpose driven and use case specific application journey is with the customer. And that is where they open up. They we make them realize that this seems to be a problem and which so so we have to go beyond their day to day operations and really have those candid conversations and then make them realize that there is a problem in their system. That is one way of, and then that is where we build smaller digital transformation enablers, we call it. So we build smaller solutions, use case specific, we solve that particular problem.
But we have to keep in mind the larger landscape. Like if I solve problem, A problem B problem C all the smaller. But the data has to be well managed at a larger scale. That is one area. And second is because this acquisition creates a challenge and process challenge.
We then came up with this, our positioning statement, as we discussed about applying data engineering properly. And then we are because we are from Microsoft Fabric. We are from Microsoft Tech stack background. I mean, the natural choice for us was using Microsoft Fabric, as a platform, which is SAS platform of engineering and that offers you, data ingestion services. And what I mean by data ingestion is imagine the same scenario again, multiple manufacturing companies being bought over. So so I can source data from company A from one CRM.
And I can source CRM data from another company. But there could be another CRM and there could be another here, and there could be another application and so on. So I can collect information from all the systems, ingest that into this platform, apply engineering, apply science, apply AI I build some process model, build some AI models, and then and then come up with nice visualization. And that is where we use Microsoft Power BI as one of the key strengths I think we have been operating with. We have built so many, amazing interactive like, dashboards, but then dashboards. This visualization is one part, but can we bring out actionable insights? That's another very important thing.
And then can we bring some conversational flavor to the way people can interact with the data? That is another third thing we were able to do. And then we were also we also explored is it just power BI or can we also build solutions in parallel where we can kind of give them ChatGPT kind of interfaces, where the end users could be business users and they can just simply chat to the data and can they get some relevant, very insightful information. And then can you start with data? Right. The data first? Yeah. We we start with conversations. If you look at my LinkedIn, backdrop, it says it all begins the conversation I see that is how it is.
And you have to converse. And that is then we convert them into discovery sessions. We convert them into UX sessions, we convert them into ideation sessions, we convert them into then process sessions, identify process chaos, data chaos, and then present them that describe the potential problems which probably needs to be solved.
That will really result into some better digital enablement. And then once they are in the mode of enablement, that is where transformation is cost. So that kind of a journey, and that is where probably we have taken a lot of time.
And I do arrive at this stage of our business. I mean, we have spent the journey. Yes, I can say that I have spent a lot of time with all our existing clients. One good thing which happened in the second orbit program is go back to your customers, focus on them first rather than acquiring new customers.
So we went back to our existing clients. We need all this nice process back to them and as a part of that, our businesses are doubling with our existing clients. So that's our one learning from that other second orbit program through which we have been able to connect to wai and the rest of team members. Also, it's really wonderful because it has been a good journey. That is what we do. So that is hopefully I am able to answer you what success? Definitely.
So we have seen at least I can say that yes, we we go way beyond so Parag and kirti. We are already at the dawn of 2025. So looking ahead 2025 and going forward, how do you see the roadmap for Parallel Minds and Datasmith AI? yeah, so year 2024. We are invested a lot in defining strategies in positioning our company. And now if you check on LinkedIn so we are 37 open positions right. So that is we all found out that position very strategically there, which kind of people we really need to grow our businesses.
Right. So data senior leadership we are looking for that is where even when we attended second orbit session, the suggestion of an Anand sir was always that whenever you want to grow, you want to scale, you cannot grow by yourself. So then you need people is ownership.
So that kind of people we are looking for to get or to get us into our journey so that they can grow and we can also work with them that is one thing. And even the second stream of , hiring we want to do is like the junior level people that we will consider, then we will grow, then we will nurture them and we want to enroll them in the possibilities of future is what because we are so focused? There is something called us paratroopers program. The paratoopers program, as I mentioned, we do a very specific hiring, a specific skill set of specific mathematical statistical background, awareness on machine learning, awareness of, in general AI as a landscape.
And then we hire them with the Python skillset and so on. And then we put them on in a hands on mode for almost two months or four months and six months. And the program is such that the so so once the engineer goes to that program, they are 100% ready for production. That is, one second is the paratroopers program's agenda is also to offer the best value to the customers.
So what we go to the customers and tell is like, this is our people. These are our people who are at six months of experience, but they are ready for production. And we also offer them something called as a fixed price for first year, fixed price for second year, fixed price for third year. So what happens is the customer is able to visualize his expense for coming three years. If you were to or if you were to, or if they were to hire a paratrooper from our side on their project in an augmented, stuffing mode, that is one second is we also know how much money we are going to make in future.
And, and, and we, we kind of tailor our program in such a way based on the customer's requirements. Although we are not a staffing company, neither we are a training company, but we want to be part of, the engineering team with the customer, but at the same time offer value not only in terms of the money, but also in terms of the quality of resources which we are providing to the customer. So that is what the paratroopers program is all about as far as in general, parallel minds strategy. You asked what 2025 looks like. Honestly, it's not just about for us.
We have arrived at a point where it's not just about 20, 25 anymore. We have transformed ourselves from parag's company, I company to then kirti's and parag's company, and then instead of that being instead of me, then we convert that into actually a we company where all the people are enrolled in the possibility of future. And we have a nice plan, which has been arrived from between 2025 to 2029. And there is a specific plan and an event which is supposed to happen in 2030. So we are now coming up with a statement called as four plus one strategy.
So what happens in coming four years and what happens in the fifth year? That is what we are looking at. So that's our horizon looking like we are looking forward. I am I can talk when I am talking in 2030, we are going to achieve a specific target. We are going to be established as a very, I would say, Microsoft fabric driven, data engineering expertise company. At the same time, we will be growing, datasmith AI into Gen AI space, not only in the application space, but in the entire ecosystem of Gen AI, including which might include services LLM development, SLM development, fine tuning of models, complete AIOps, complete MLOps, or rather complete Gen AI ops. That is something which you want to establish and we want to cater into very specific areas, be it oil and gas or be it real estate or be it healthtech, and be manufacturing only for majorly.
That is what we need to do as we intend to do. As far as Gene AI is concerned. And our custom software development, which we now rebranded to call to make calling to be called as Product Engineering Services. That is something which we are going to expand by replicating our the way we work with existing clients, and primarily we work with our existing clients to really add more value to their businesses.
So that is what our future current and the latter future looks like. Looks exciting. Wonderful. Thank you.
Thank you so much for being here, Parag. And Kirti it has been such an insightful conversation. We would definitely want to have a sequel to this conversation where we talk about more in detail about more, areas that Gen AI is going to touch. Correct. And it has been a pleasure hosting you.
It has been our pleasure and thank you, thank you, thank you very much. Thank you so much. Viewer’s for watching this podcast. It has been such an enjoyable and insightful conversation with Parag and Kirti Bihade and there from Parallel Minds and Datasmith AI. And stay tuned for our next podcast. Thank you.
2025-01-29 20:53
