Field Notes
During my internship, I worked on identifying real-world problems and exploring whether they could become viable startup ideas. A major learning was understanding that finding the right problem is often harder than building the solution. I learned firsthand how powerful Reddit and online communities can be for problem discovery - observing recurring complaints, workarounds, and discussions to understand what people genuinely struggle with. I also applied principles from The Mom Test: asking about people's actual past behaviour and experiences rather than pitching an idea and asking whether they liked it. This helped me distinguish strong problem signals from polite agreement. On the technical side, I explored JEPA models and AI architectures, along with rapid prototyping of different ideas. Domains explored: - ML & Generative AI - experimenting with generative architectures and understanding where they create genuine user value vs. where they are solutions in search of a problem. - Medical AI - exploring pain points in patient-doctor communication, diagnostic support, and health data access in the Indian context. Overall, the internship taught me problem discovery, user research, validation, asking better questions, identifying genuine pain points, rapid experimentation, and knowing when to pursue, iterate on, or discard an idea. Motto: Beyond every horizon.
Notes & Reflections
“"Existence precedes essence." — Jean-Paul Sartre”
— S Anuj Srijan
Domains
Margin Note
My dream isn't simply to build a successful startup; it is to find a problem that genuinely matters to people and build something that solves it well.
Also in ABC 2026
Vigil
Cross-source signal correlation layer for manufacturing maintenance — correlates alarm logs, sensor trends, and operator observations in real-time to flag developing failures before they happen.
Influex
Creator intelligence platform bridging the gap between creators and brands — data-driven insights, authenticity assessment, portfolio management, and campaign performance tracking.
SpikeNest
Explored industrial sustainability and methane emission management, predictive maintenance using spiking neural networks, and attitude adjustment for CubeSats.
AgentOps
Explored AI-driven agentic solutions across Finance, Real Estate, EdTech, and MSME verticals — focused on identifying meaningful problems and validating real user pain points.
EdgeDaemon
Using AI to reconstruct, analyze, and improve decision-making processes — taking past decisions, reasoning, and outcomes to reconstruct how a decision was made and identify areas for improvement.
VoxNova
Explored AI reliability across Insurance & claims processing, Customer Service, Geospatial vision, and Warehouse logistics — mapping where AI breaks and how to make it trustworthy.
EquationX
Explored the intersection of Cybersecurity, AI & ML, and Network Security — building intelligent systems that detect and respond to network threats.