AI Product Manager
I build AI products and
ship them end to end.
I'm Bhaskar. I spec the product and build it myself, from discovery to production.
The Thesis
The hard part of an AI product is never the prompt. It is the safety, the grounding, and the judgment that never make it into the demo.
That is the part I build for.
Right now
What I'm building today
Shipping AI products with safety built in from the first slice, and staying close enough to the tools that my roadmaps match what can actually ship.
Currently exploringGoing deeper on evals and guardrails for agentic AI workflows.
Featured work
Three builds that prove it
The ServiceNow Co-Pilot, Job Agent, and the Dental Co-Pilot. Three AI products taken from idea to shipped. The full set is in the grid below.
Where it's going
What I'm learning, and what people say
The next problems worth solving, and the clients and teammates who have worked with me on the way.
The ask
Let's build something
If you are building with AI and want a product person who can also build it, the door is open.
Built for
- ServiceNow
- The Dental Specialists
- Mark My Zone
- AirAsia Move
- Institute for Dental Implantology
- The R Group
- OnLoop
- Wheedle
- Moment It
Real clients, real production systems.
Built with
The stack behind the work
Tools I build with, not products I sell.
- Next.js
- React
- Supabase
- Vercel
- GitHub Actions
- Webflow
- Claude / Claude Code
- LangGraph
- LangChain
- RAG + LLMs
- n8n
- Apify
- Runway
- Kling
- Sora
- ElevenLabs
- HeyGen
- AnimateDiff
- Figma
- Notion
- Google Analytics
- Linear
All work
Everything I have shipped
Selected work, built end to end. Each opens its full case study.
About
The person behind
the builds
I am an AI Product Manager who ships what I spec. I stay close to the tools, LLMs, RAG, and coding agents like Claude Code, so my roadmaps match what can actually ship.
The clearest example is an AI copilot I shipped for a multi branch dental group. It answers only from the clinic's own records, strips patient identity, cites every claim, and runs under hard cost caps.
So an unsafe answer is structurally hard, not just discouraged. I shipped it in six independent slices, each cleared by code review, a security pass, and a full test suite.
I have been building and operating since long before it was AI. I ran rice export logistics across the Middle East and West Africa, then started a restaurant and a generators venture.
In 2022 I founded Mighty Minds, an AI product and growth studio: 50+ projects, a team of 4. I grew an audience from 0 to 54.7K with zero ad spend, and stopped an AI voice agent when the data said no.
Currently exploring: going deeper on evals and guardrails for agentic AI workflows.
Bhaskar · AI Product ManagerTestimonials
What people say
Bhaskar is a strategic webflow expert with a deep understanding of user experience. He excels at understanding target audiences and crafting impactful strategies. … an invaluable asset for any startup looking to boost their online presence. I highly recommend Bhaskar's services.
Contact
Let's build something.
If you are building with AI and want a product person who can also build it, I am easy to reach.
Email me