About me

I look for the system underneath repeated effort.

I am an AI Product Manager who started as a developer and learned product by building and running companies. I stay close to users, technical teams, and the operating realities that determine whether a product becomes useful.

Developer · Founder · Hands-on product manager

My point of view

Technology becomes a product only when it fits the user, the operating environment, the economics, and the consequences of being wrong.

I learned this by moving between code, customers, teams, model behavior, delivery systems, and business constraints. The work I enjoy most starts with ambiguity and ends with a product—and a team—that can keep learning.

My journey

Each chapter expanded what I believed product work had to include.

Developer

Learning what happens after the code works

I began by writing software. Seeing some products become useful while others went untouched made me curious about the decisions before development: whose problem are we solving, what evidence do we have, and what has to be true for this to work?

Founder

Building SochWare

I co-founded a technology company in Nepal, helped grow it from four to more than forty-five people, and contributed to 45+ products across agriculture, healthcare, education, enterprise, and civic services.

Accessible AI

Leading E-Agrovet

I led an AI product that helped farmers identify crop disease from a smartphone photo and connect with expert support. The work won the Microsoft Imagine Cup 2018 World Artificial Intelligence Award.

AI operations

Learning from human-in-the-loop systems

At CloudFactory, I led computer-vision annotation pilots and a 25-person delivery team. I learned how data quality, review design, tooling, and operating discipline shape model outcomes.

Product systems

Rebuilding execution

At Yarsa Tech and YHH, I worked on the system around the product: discovery, decision rights, roadmaps, P&L, delivery rhythm, commercialization, and succession. At YHH, 164 interviews informed two commerce platforms that increased activation by 37% and first-quarter revenue by 40%.

Applied AI

Accessibility, agents, and trustworthy evaluation

At USF, I built accessible audio AI, a research agent with structured outputs and human review, and an evaluation instrument for LLM sycophancy. I also helped deliver cybersecurity learning to 1,500+ Florida government employees.

Selected proof

Different domains. One habit: learn the system before changing it.

45+products built or launched

4 → 45+company growth

92%accessible AI field-testing accuracy

37%activation increase at YHH

1,500+government learners reached

≈1,400LLM evaluation responses

Where I am going

Hands-on AI product work where I can make visible change.

I want to join a startup or mission-driven team where I can stay close to users, collaborate deeply with technical teams, and help shape both the product and the way it is built. I am especially interested in AI products where reliability, human judgment, access, and responsible deployment matter.

Download my CV

Continue to the evidence

The story is context. The work shows how I think.

Explore product transformation, accessible AI, trustworthy evaluation, and practical AI built for real operating environments.

Explore my work