Product thinking
When to Sit in on Discovery, Not Just the Demo
Getting closer to customer reality, and knowing when to listen before deciding what to build.
About · Kathmandu, Nepal
Product manager,
founder, builder.
I'm Prakash Poudel Sharma. I work independently, helping teams make product decisions and adopt agentic software development—from assessing codebase readiness to setting up the workflows around it.
Here, I write about product thinking, engineering with AI, and the systems I explore hands-on. Building, asking questions, and writing things down are how I make sense of the work.
Independent work. Over a decade building software.
What I work on
From product decisions to software that ships.
I bring product and engineering experience to freelance PM engagements and consulting with teams adopting coding agents.
Freelance product management
Discovery, prioritisation, and delivery. I help teams clarify the problem, decide what matters next, and connect product decisions to the work being shipped.
Codebase readiness audits
Assess how well coding agents can understand and work in your repository: its structure, documentation, tests, and feedback loops. Identify the changes that would make agent work easier to review and verify.
Software factory setup
Connect requirements, coding agents, code review, testing, and release into a repeatable workflow. Make responsibilities and checks explicit so the team can follow work from an idea to a verified change.
Agentic coding adoption
Help teams choose where agents fit, establish working practices, and learn through real development tasks. Build the team's ability to steer, review, and improve the workflow over time.
How I got here
Different roles.
A builder throughout.
I founded Truemark Technology, a product engineering studio. That work put me on both sides of the conversation: building software and taking responsibility for the team and business around it.
Earlier, I worked in software engineering at Leapfrog Technology. I've also worked as an independent contract software engineer for Luthor (YC F24).
Today, those experiences meet in my independent work. I care about choosing useful problems, making the code understandable, and giving teams a way to check that what they ship does what they intended.
More about my experience →What I write and explore
A notebook for the work—and the curiosity around it.
Some essays start with product discovery or leadership. Others follow coding agents into real codebases. My homelab gives me room to experiment with infrastructure, music, and systems for their own sake.
Engineering with AI
Exit 0 Is Not Evidence: Why AI Coding Agents Need Receipts, Not More Memory
What it takes to trust an agent's work: reviewable changes, meaningful checks, and evidence.
Hands-on experiments
The Homelab Music Recommendation Field Guide: Every Signal, Trade-off, and Winner
Exploring music discovery at home, from metadata and listening habits to the trade-offs behind recommendations.
For a longer read: Agent-ready React explores codebases that agents can work in; The Parallel Developer covers working across agents and workstreams.
Browse all writing →Work together
What is your team working through?
If you need a freelance product manager, a clearer picture of your codebase's readiness for agents, or help putting agentic development into practice, I'd like to hear about it.
Let's talk about your project