@sherinjosephroy
Co-Founder at DeepMost AI • Building contextual AI systems • Writing honest notes on enterprise AI, products, and entrepreneurship
Sherin Joseph Roy Co-Founder & Head of Products at DeepMost AI, Bangalore, India. What I'm Building At DeepMost AI, I'm focused on solving the trust gap in enterprise artificial intelligence. We build AI systems with contextual intelligence—technology that understands business context, organizational constraints, and when to admit uncertainty rather than providing confident but unreliable answers. What I Write About Through "Notes by Sherin," I share:
The reality of AI adoption in enterprises (beyond the hype) Product development in rapidly evolving markets Honest reflections on building a company in the AI space Human-centered approaches to technology development Lessons learned from both successful and failed implementations My Approach
I believe the most valuable AI systems will be those that augment human judgment rather than attempt to replace it. Technology should make us better thinkers, not do our thinking for us. This philosophy shapes everything we build at DeepMost. Background Before co-founding DeepMost AI, I spent years in product development and system architecture. I combine technical depth in AI with practical understanding of how organizations actually adopt and use technology. A Note on Names My legal name is Sherin Roy. Joseph is my baptismal name, which I use consistently across all professional contexts. Both names refer to the same person—me. Connect
Website: sherinjosephroy.link Company: DeepMost AI https://deepmostai.com Location: Bangalore, India
Writing for builders, founders, and anyone interested in creating technology that's meaningful over impressive.
Beyond my professional work, I remain actively engaged with the broader technology community through open source contributions, technical advisory relationships with early-stage companies, and participation in industry discussions about responsible AI development and deployment practices. I welcome connections from others working on enterprise AI challenges, technical leaders navigating similar implementation journeys, and anyone interested in honest discussion about what building production AI systems actually requires beyond the simplified narratives that dominate much industry discourse.
Feb 26 · 5 min read · Kerala records over 40,000 road accidents annually. SurakshaNet is a real-time monitoring platform I built that uses 13 AI agents across 3 clusters to assess road risk, attribute causality, and trigge
Join discussion
Nov 14, 2025 · 4 min read · Hey there—whether you're a dev tweaking LLMs for your app, a marketer crafting AI copy, or just curious about squeezing more juice out of tools like Grok or Claude, prompt engineering is your secret sauce. At its core, it's the art (and science) of c...
Join discussion
Nov 14, 2025 · 5 min read · Hey folks, it's been a wild ride this year. Remember back in 2020 when everyone was scrambling to learn React like it was the holy grail? Fast-forward to 2025, and I'm sitting here, sipping my third coffee of the morning, staring at a codebase that's...
Join discussion
Nov 5, 2025 · 3 min read · When I first started experimenting with enterprise AI, I made the same mistake everyone else did —I tried to make the model smarter. More parameters, better embeddings, new APIs, fine-tuning loops… all of it.But no matter how much power I threw at it...
Join discussion
Oct 23, 2025 · 5 min read · Look, I've tried a lot of AI coding tools. GitHub Copilot, Cursor, all the usual suspects. They're fine. They autocomplete my code, sometimes they get it right, sometimes they don't. But Kiro? It's different in a way that took me a while to fully app...
Join discussion
