Forward Deployed Engineer

Tuning Research (WorkableAI)

Mumbai, Maharashtra, India

Forward-Deployed Engineer / AI Product Manager

Tuning Research

India (hybrid, at client sites)

About Tuning Research

Tuning Research is an applied AI engineering company, founded by IIT Kanpur alumni, that builds production-grade agentic AI systems for enterprises. We’re a 20-person team that works as an extended AI team for our clients across knowledge and documents, business operations, video intelligence, voice AI and marketing automation.

About the role

We’re hiring a Forward-Deployed AI Product Manager to own client transformations end to end: scope it, design it, build parts of it, ship it, and get it used.

You’ll be the one person from Tuning Research who lives inside the client’s business. Typically that’s a mid-to-large Indian enterprise replacing spreadsheets and legacy systems (HRMS, ERP, accounts, operations) with AI-native software. You sit with their teams, work out what should be built, design it, build and ship parts of it yourself, and make sure hundreds of people actually use it. A small remote team of AI, full-stack and mobile engineers works with you, and our founders stay closely involved.

The title says product manager, but the job is closer to a forward-deployed engineer: half product, half engineering, all of it on the customer’s ground.

What a week looks like

Monday: At the client’s office. Stand-ups with business teams to agree priorities and blockers.
Tuesday: Building. Backend endpoints, data migration scripts, an AI extraction flow, or a Figma screen.
Wednesday: On the floor or at a site. Watch people use what shipped, demo it, capture what’s wrong.
Thursday: Review the engineers’ code, test on real data, deploy, and fix what breaks.
Friday: Write next week’s specs, then review with our founders and client leadership.What you’ll do

Scope the right thing

Run discovery across departments and map how work, material and money actually flow.
Separate what people ask for from what they need. Decide what is configuration, what is a one-off, and what becomes a product feature.
Turn findings into a prioritised monthly scope with clear acceptance criteria, and say no to things that would fragment the platform.Own the client relationship
Be the single point of contact for everyone from frontline supervisors to the MD.
Run weekly working sessions, fortnightly reviews and monthly scope meetings, and keep one tracker of every request and decision.
Explain trade-offs, delays and risks plainly, before they become surprises.Design the system
Own the system design: the data model, integrations with existing ERPs and accounting systems, and roles and permissions.
Decide where AI belongs (document reading, matching, alerts, copilots) and where plain software is enough.Build, verify and deploy
Write backend code yourself: APIs, integrations, and data migration from Excel and legacy systems.
Review every pull request from your engineers, and test on the client’s real data before anything ships.
Run releases and own incidents and fixes in production.
Set up evals for AI features so accuracy is measured, not assumed.Design the experience and drive adoption
Design screens in Figma for people who don’t type much: offline-first, in Hindi and English, for low-cost Android phones.
Test designs with real users before the team builds them. Train users and write simple guides.
Watch usage, find where people drop off, and fix it.What you need
3–6 years building and shipping software, including at least 2 years in a client-facing, founding-engineer, product-engineer or forward-deployed role.
Strong backend skills in Python or TypeScript/Node: REST APIs, PostgreSQL or similar, and SQL you write yourself.
Hands-on experience with LLM applications: prompt design, document extraction, retrieval, tool-calling agents, and evaluating output quality.
Able to read, review and reject code. Comfortable with Git, CI/CD, Docker and one cloud (AWS, GCP or Azure).
Product sense: you can write a clear spec, prioritise with explicit trade-offs, and design a usable screen in Figma.
Able to run a workshop with a site supervisor in the morning and an MD in the afternoon, and you write plainly.
Fluent Hindi and English.
Willing to work from client offices and travel to sites, daily in the first month or two of an engagement.Nice to have
You’ve built ERP, HRMS, accounting or operations software, or you know Tally, GST, TDS or payroll.
Experience with construction, infrastructure, manufacturing or field-operations businesses.
You’ve shipped a mobile app (React Native or Flutter), ideally offline-first.
You’ve led data migration from Excel and legacy systems.
You speak a regional Indian language beyond Hindi.Who succeeds here

You have high agency and low ego, and you’re comfortable when the plan changes. You’ve shipped something real, whether at a startup, a hackathon or on an on-call rotation, and you’d rather fix the problem than write about it. You stay calm while a client watches you fix something live.

Work setup

Location: Hybrid, at client offices in India. Daily for roughly the first 1–2 months of an engagement, then about 3 days a week.
Travel: Regular visits to client sites.
Team: You lead a remote AI engineer, full-stack engineer and mobile engineer.
Reports to: Tuning Research founders.
Engagements: Typically 12-month programmes, with the chance to roll out across the client’s group companies.What success looks like
30 days: You’ve met every department and mapped how work and money move. The first monthly plan is agreed and the build is under way.
90 days: The first module is in use at pilot sites, the engineers ship every week, and the key build-vs-integrate decisions are made with the client.
6 months: Core modules are live and client leadership treats you as part of their team.
12 months: One platform runs the client’s core operations, and leadership sees the numbers daily.How we hire
Intro call (30 min): Your background and why this role.
Discovery role-play (60 min): One of us plays a client’s department head describing a messy problem. You run the conversation and play back what you’d build.
Take-home (one weekend): Build a small API that uses an LLM to extract a real-world business document into structured fields. Add a one-page spec and one Figma screen for the end user.
Code and design review (60 min): Walk us through the take-home, and we review a pull request together.
Founders’ conversation: Working style, travel and expectations.

Fill out the form https://forms.gle/L9M6MDArXLP6KFFc8

Score my resume against this job, free →