Founding AI Engineer

Terasol Technologies Pvt. Ltd.; Mumbai; On-site; INR 1,600,000 - 2,600,000; Full time

Job summary

Terasol Technologies Pvt. Ltd. in Mumbai is seeking a founding Product Engineer, Applied AI to bridge our multi-agent AI system and client reality. You’ll own outcomes end-to-end, from discovery to live deployment, tailoring AI agents to each client’s data, brand rules, and compliance.

You’ll work closely with the CTO, Codex/Claude Code, and client stakeholders, ship production code, and build scalable customization patterns as we grow from 10 to 100 clients.

Qualifications

• Real engineering chops: read code, write production code, debug agent failures.
• Comfort with agents as a system: understand prompts drift, context engineering, tool use vs fine‑tuning.
• Client-facing instincts: communicate with CMOs, IT heads, and brand managers.

Responsibilities

• Customize agents: tune prompts, build capabilities, configure workflows, wire integrations.
• Own client outcomes end‑to‑end from discovery to live deployment.
• Problem-solve in real time: diagnose and redeploy to meet client needs.
• Handle kreative feedback and assess prompts vs brand guidelines.
• Feed learnings to core product and grow customization patterns.
• Build playbook for scaling customization across clients.

Skills

Engineering chops

Agents as a system

Client-facing instincts

Creative literacy

Shipping bias

Job description

We’re building in a market that hit $47B in 2026 and is on track to hit $107.5B by 2028 and we’re not chasing that opportunity from the sidelines. Our founding team brings marketing expertise from Kotak and Amazon paired with deep AI chops, we were selected as one of just 20 startups by Google, and we already have paying customers like Kotak and Myntra using the product in production. Join us early and help build the infrastructure that turns AI marketing adoption (already at 88%) into real, scaled results for customers who are already paying for it, not case studies.

Founding Product Engineer, Applied AI

Location: Mumbai • Hybrid • Reports to: CTO

WHAT THIS ROLE ACTUALLY IS

We’ve built a multi-agent AI system for marketing. It works well out of the box but every client (their data, workflows, brand rules, compliance constraints, output preferences) is different. Generic doesn’t ship results. Customised does.

This role is the bridge: you sit between our agent architecture and the client’s reality. You understand both deeply, and you make the agents do what each specific client needs them to do.

Think Palantir Forward Deployed Engineer, but for AI agents. An engineer who owns product outcomes end‑to‑end.

WHAT YOU’LL DO

Day‑to‑day split: roughly 60% in Codex and Claude Code shipping customisations, 40% in client rooms understanding what to ship next.

• Customise the agents. Tune prompts, build new agent capabilities, configure workflows, wire up integrations. Codex and Claude Code are your primary IDEs. You ship working code, not specs for someone else to build.
• Own client outcomes end‑to‑end. From discovery to live deployment, you are the person responsible for whether the agents actually solve the client’s problem. No PM‑engineer handoff to hide behind.
• Problem‑solve in real time. Client says “the output tone is off for our brand” or “the research agent is missing X data source” — you diagnose, fix, and redeploy, usually within days.
• Hold your own on creative. When a CMO says “this doesn’t feel right,” you can question intelligently — is it the prompt, the input brief, the brand guidelines themselves, or the client’s own confusion? Most engineers reach for the prompt. The good ones diagnose first.
• Feed learnings back to core product. Anything that gets customised 3+ times becomes a platform feature. You decide what graduates.
• Build the playbook. As we scale from 10 to 50 to 100 clients, the customisation patterns you create become the foundation others work from.

WHAT YOU NEED

• Real engineering chops. You can read a codebase, write production code, debug agent failures, and ship to live clients. AI tools are your accelerant, not your crutch.
• Comfort with agents as a system. You understand how LLMs fail, why prompts drift, what context engineering means, when to use tools vs. fine‑tuning, why eval harnesses matter. If “MCP,” “tool use,” and “agent harness” feel like jargon, this isn’t the role.
• Client‑facing instincts. Comfortable in a room with a CMO, an enterprise IT head, and a junior brand manager — and you can switch register for each in the same meeting. If client calls feel like tax to you, don’t apply.
• Some creative literacy. You don’t need to be a designer. You do need to be able to tell the difference between “this output is technically broken” and “this output is technically fine but creatively wrong” — and chase the right thread.
• Bias toward shipping. A working hack today beats a clean design next quarter.

YOU’LL THRIVE HERE IF

• You’re an engineer who realised products live or die on customisation, not architecture.
• You’ve been frustrated by being kept away from customers.
• You like building with AI tools and want a role where that’s the job, not a side experiment.
• You like marketing as a domain — or you’re willing to fall in love with it.

YOU WON’T THRIVE HERE IF

• You want to optimise systems in a corner with headphones on.
• You think client calls are below your pay grade.
• You think AI is overhyped (we don’t).
• You need a fully scoped ticket queue and weekly sprint rituals.