Hi, I'm Nilesh.

I build systems that are simple, scalable, and grounded in real-world needs. My work sits at the intersection of AI infrastructure, agent systems, and product engineering—always with a focus on practicality over complexity. I’ve built for Healthcare, Gaming, Supply Chain, and Finance.

I care deeply about usability. The tools I build are intuitive and designed with their users in mind. Whether it's code, systems, or teams, I work to leave everything better than I found it.

Recent Work

• Pollen.tech: Led ML team, designed and deployed the in-house multi-agent system with 3 named agents in the domains of demand prediction, pricing, and sales strategy. Did all adjacent data engineering, MLOps, backend stuff before growing the team.

• Sizzle.gg: Led GenAI migration for video processing, built scalable recommendation engines, contributing to platform growth from 10K to 100K+ users.

Projects:

• MIT Media Lab : Worked with MIT Nanda MCP initiative on technical architecture as well as industry partnerships. See a live demo

• ICAR, Govt of India: Helped a PI and PhD students build an ML tool using image recognition. Farmers could take a pic of their cattle to get an early heads-up on health problems.

• WarehouseAI : Helped a startup build an MVP from 0-1 on contractual basis. Did design, frontend, backend, ML, security, DevOps etc. preview here, ping me for ML demo

What I Believe In

Technical:

• Avoid over-engineering.
• Avoid abstractions unless they pay for themselves in readability or leverage.
• Simple is maintainable. Maintainable is scalable. Scalable, sometimes, becomes elegant.

General:

• Picking the right problems to work on is higher leverage than perfect execution on the wrong ones. Beware of busy-work.
• Do what you said you'd do, when you said you'd do it. Underpromise and overdeliver — reliably.
• Know your audience. A producer's output is meaningless if the consumer cannot parse value from it.
• Have pride in your work, and aim for internal balance: discipline without burnout, ambition without tunnel vision.

Reading List

• Ilya 30u30
• dl course
• dl book
• dl cs 294
• losh
• vending
• dia tts
• abyss essay
• pre training
• air travel planning (Google)
• stripe atlas
• o1 essay
• saas hack?
• cool Google paper
• intrep | gemma
• in the plex
• cpu talk
• markov chain
• blue zones?
• machines of loving grace
• leopold paper
• timeline post
• timeline essay
• cool AWS DB video
• coding system prompts