Why domain expertise now beats coding for building products

Why domain expertise may matter more than coding now

Muzamil Hasan opens by framing the episode around a shift he sees in AI. Technical people are already using AI to make their work easier, and some are training systems that could eventually replace parts of their own jobs. For non-technical professionals with domain experience, he argues, the change matters even more. He describes the old barrier of needing a computer science background to build software as something that has largely disappeared. “I would argue that the problem right now should be for computer science people,” he says, because people with experience in other fields can now turn that experience into products quickly.

Introducing Shayan Baig and AlphaGenPro

Muzamil introduces Baig as the founder and CEO of AlphaGenPro, which he describes as a PSX research platform. He notes that Baig has spent roughly eight years in trading and the stock market, and that AI made it possible to build the tool he had wanted for himself and then extend it to other people.

Baig thanks Muzamil and explains why he accepted the invitation. The product took about three years to build, he says, and once it was finished the team realized marketing should have run in parallel. He says such products are “very rare,” and that stock market advice in Pakistan has long been dominated by tips.

From electrical engineering to stock market research

Baig says he trained as an electrical engineer. The economic shock of COVID, he explains, made him want another source of income, and that search led him through digital marketing and eventually to the stock market. He also completed a master’s in engineering management, which included finance subjects.

The problem he saw was the quality of information. Tips were common and quality information was scarce. As awareness grew, more institutes and educational resources appeared. Over the following eight years or so, Baig says, he worked through the difficulties, summarized what he had learned, and built a top-to-bottom model with a co-founder.

The fundamentals and technical logic behind the model

Baig’s core premise is that share prices “always follow improving fundamentals” in the companies behind them. He points to COVID, when most markets fell but some stocks held up or rose. Internationally he names Zoom and Netflix, and locally a company he calls Oxygen Pakistan. That pattern led him to ask what was behind those companies’ performance.

He traces the Oxygen example to hospital oxygen demand during the pandemic. Higher demand lifted supply and profitability, and that later showed up in the share price. He also describes Systems Limited, whose price he says rose from around 80 rupees to 800 as work-from-home and digitization took hold. He presents these cases as turning points that convinced him the link between fundamentals and price could be captured systematically.

Baig explains that the model covers both main schools of stock analysis. Fundamental analysts look at balance sheets, income statements, and cash flow statements, while technical analysts look at trends. He says the model combines the two, and that in his view no model defines the technical side in this way.

Turning qualitative signals into scores

A recurring theme is the conversion of qualitative information into numbers. Baig says the team captured qualitative information as quantitative scores, so that factors such as sector conditions and company performance could be compared consistently.

He illustrates this with refineries. The Russia-Ukraine war, he says, pushed up refinery margins and the prices of products like diesel and motor gasoline, which lifted profitability and showed up in results. He says the model captured this quickly. On top of that, the model layers macroeconomic and geopolitical context, asking how events affect the country, which sectors, and which companies, before examining each company’s technical trend.

Baig frames the output as research rather than advice. He says earlier methods such as web scraping raised ethical concerns, and that the current approach uses AI agents that navigate sources, extract data into a dataset, and pass it through another AI tool before presenting it in an interface designed so users are not overwhelmed.

Walking through the product

Muzamil asks Baig to demonstrate the platform. On the home page, Baig shows a demo portfolio of PSX companies and explains that the platform tracks each holding’s price movement and related news. He points to insider buying at Millat Tractors, and to a monthly sales report for Indus Motors showing 4,000 units sold. He presents these as the kinds of developments the model surfaces automatically for each user’s holdings.

The feed shows upcoming dividend book closures and results announcements, and can be filtered to the companies a user holds or watches. Corporate actions, insider activity, and material information from the exchange appear in one place. Baig describes the portfolio as interactive, meaning it flags negative trends and shows how far a holding has fallen over a given period, using one stock that he says has been falling for 59 days as an example.

Next, Baig walks through market pulse and compare tools. The market pulse starts with the KSE-100 index as the benchmark, then moves down to sectors and individual companies. He uses the refinery sector as an example, noting a company whose price rose from 38 to 100. A compare tab places several companies side by side against a benchmark. On a company page he shows a business score, a technical status, and a fair value estimate, and notes that fair value is “quite subjective” and depends on assumptions.

Building, testing, and monetizing the product

Baig says the product was developed over roughly four months by two people, himself and his partner, drawing on about three years of earlier work in Excel and Power BI. The shift to AI, he says, let the team move from partial automation to a more holistic system that included qualitative information.

Muzamil asks whether there are paying customers. Baig explains that the platform was offered free for four months and collected feedback from more than 2,000 users, some of whom asked for features that were then built. When costs reached a level the founders could no longer cover from their own salaries, and after the free period had shown users how the product worked, they began charging last month. Baig says they have around 100 users within the first month, and that marketing has not yet been a major focus.

He also reflects on why many builders struggle. Many people, he says, design a product in their head, build it, then face a paywall. He argues for a feedback relationship with early users before monetizing. He stresses that the product came from his eight years of finance experience, not from AI knowledge, and that AI helped him build what he already understood.

Using AI tools and the question of reliability

Baig says the team used GPT, Claude, and Gemini, among others, and that the platform connects to an MCP server so that Claude can read the portfolio and make updates. He describes a feedback loop in which the model can identify why a score failed to predict a stock’s performance and adjust over time. He repeatedly stresses that the product is a research platform, that users should treat it with caution, and that they should add their own research.

Muzamil raises a broader point about Pakistani business culture. He observes that people often dismiss local products quickly, listing reasons they will fail, instead of testing them and giving detailed feedback that helps improve them. Baig agrees that feedback from real use is what moves a product forward, and encourages viewers to test the platform, check its claims, and share their experiences in the comments.

Closing thoughts on AI, timing, and access

Toward the end, Baig reflects on his own experience as a young investor, saying it was hard to enter the market and that newer tools are making access easier for younger people. Muzamil ties the conversation back to his earlier argument, comparing AI adoption to digital media in its early years, when early adopters often gained an advantage. Baig agrees that early adopters of AI are likely to benefit.

The episode closes with Muzamil asking viewers whether they trade in the Pakistani stock market, and inviting them to share their experiences in the comments.