Website:
getsketch.ai
Job details:
Technical Product Manager – AI-Driven Web Applications
Company: Canvas (canvasx.ai)
Location: Remote
Experience: 2–4 Years or more
Compensation: INR 15–22 LPA + ESOPs
Please note that failure to submit the required assignment mentioned in the application form will result in the rejection of your application.
THE SHORT STORY
We're Canvas — a research and R&D lab building AI products. We started the day ChatGPT launched and have shipped more than 15 AI products since. Right now we're building Getsketch.ai, and that's where this role plugs in.
You'll work directly with our CEO and CTO, owning the product layer of Sketch and whatever ships next. Small team, fast loop, real product decisions made in the room — with you holding the pen.
THE MISSION
You'll own product outcomes end-to-end, not feature lists. The brief might be one line in Slack. Your job is to figure out what's worth building, write the spec sharp enough to ship from, get it through engineering, then watch how people actually use it — and decide what comes next.
We're a startup, and we need a technical PM. Someone who can read the codebase, sit in a Linear ticket, and still zoom out to the strategy call an hour later. If you want a backlog handed to you, this isn't the role. If you'd rather sit in pure roadmap decks without ever opening a PR or tracing an API call, it isn't either.
WHAT YOU'LL DO (THE IMPACT)
Find the Edge: Notice what's missing or broken in our products. Talk to users. Watch session recordings. Pull the data. Decide what's worth a sprint and what isn't.
Write the Spec: Turn fuzzy problems into clear specs engineers can ship from — with the edge cases mapped, the success metric named, and the cuts already made before the kickoff.
Make the Call: When the trade-off is ambiguous — speed vs. polish, scope vs. timeline, depth vs. reach — you decide. We'll back you.
Ship the Loop: Drive features from kickoff to launch. Unblock engineering, run the QA pass, write the changelog, push the release. Own the date.
Read the Signal: Set up the dashboards. Define the metrics that matter. Know within a week of launch whether a feature earned its keep, and have the call ready on what to do about it.
Speak Engineer: Sit in the architecture conversation and add to it. Understand the stack well enough to know when "two days" actually means two weeks — and when it doesn't.
Hold the AI Layer: Get deep on what LLMs can and can't do. Shape prompts, tune flows, define eval criteria, and decide where AI earns the UX and where it gets in the way.
THE STACK YOU'LL WORK WITH
Frontend: Next.js, React, TypeScript, Tailwind Backend: Node.js, Fastify, REST/LLM APIs, MongoDB AI: OpenAI, Claude, Gemini DevOps: Docker, AWS (Lambda, ECS), GitHub Product: Linear, PostHog/Mixpanel, Notion, Figma
You don't need to write production code. You do need to read it, reason about it, and know what's expensive vs. cheap.
WHO YOU ARE
The Experience: 2–4 years shipping software products. Bonus if some of that was as an engineer who moved to PM, or as a PM who's stayed close enough to the code that engineers respect your specs.
The Foundation: Comfortable reading TypeScript and tracing how a request moves through a stack. You don't need to ship the PR, but you can review one and ask the right question.
The AI Curiosity: You've built with LLM APIs — even on the side. You have opinions on what's hype, what's real, and where the next year goes. You know what a good eval looks like.
The Taste: You notice when a product is bad. At the click, the copy, the empty state, the latency. You have strong opinions about what makes software feel good — and you can articulate why.
The Judgment: Comfortable making calls with incomplete information. Comfortable being wrong, owning it, and adjusting. Allergic to roadmaps that look like wish lists.
The Craft: Sharp writing. Sharp thinking. Ownership over outcomes, not output. Willing to dig into a customer ticket, a SQL query, or a Figma file in the same afternoon.
THE SELECTION PROCESS
- Application: Fill out our application form with your details.
- The Assignment: Complete the assignment shared in the form (7 days to submit). The deliverable is a short product note, not a deck. Applications submitted without the note will not be considered.
- Note Review: We review your note for thinking, structure, and craft.
- Final Conversation: A discussion with the team to walk through your note, your process, and mutual fit.
WHY YOU SHOULD JOIN
- Your specs ship into real AI products customers use, not internal experiments.
- Work closely with the CEO and CTO. Your calls shape the product. Propose it Monday, ship it by Friday, learn from it the week after.
- Learn the craft of running production AI products end-to-end — from problem discovery to LLM evals to launch.
- Competitive compensation with meaningful ESOPs — real skin in the game.
Click on Apply to know more.