Member of Technical Staff (founding team)

Most personal assistant products hand you a chatbot and call it done. faff is built on a different premise: purely agentic solutions fail at the messy, real-world tasks people actually need help with — grocery, bills, travel, vendor calls — because those tasks depend on external services, live across calling, WhatsApp, and email, and break in ways no model handles cleanly on its own. So faff keeps a human in the loop, exactly where the automation cannot hold.

The team is small, the stack is genuinely hard, and the automation layer is being built in real time. Three engineers, six-day weeks, and a clear near-term target: 80% workflow automation in three months. A consumer product with global ambitions, built from Bengaluru.

faff's automation layer sits at 40% today. The goal is 80% in three months, and this is the seat that gets it there.

You'll join a three-engineer team as an individual contributor, own your projects end to end, and ship to production — not to staging, not to a demo environment. The stack covers browser agents, voice agents, multi-agent orchestrators, and memory systems that hold more context than a typical LLM window. None of it is solved territory. Production constraints, real users, and edge cases that no tutorial covers are the daily reality. The first week looks like the rest: pick up a problem, move fast, ship something real.

The 40-to-80 jump. The automation layer is at 40% today; you own the path to 80% in three months, across every workflow that still runs manually.

Browser agents in production. Playwright, browser-use libraries, reverse-engineered consumer surfaces — you build the automation that survives IP blocking, rate limits, and the third-order edge cases that only show up when real users hit it.

Voice agents beyond vanilla deployment. Latency at the millisecond level, model evaluation, open-source versus closed-source cost tradeoffs — you tune the system until accuracy and speed hold under real load.

Multi-agent orchestration. You design and own the orchestrators that coordinate agents across calling, WhatsApp, email, and external vendors — systems that handle the edge cases that purely agentic solutions cannot.

Memory systems with real depth. Context that exceeds a typical LLM window, built to hold the full complexity of a user's ongoing tasks — not a simple chat log.

Your own projects, end to end. You pick up a problem, scope it, build it, and ship it. No handholding once the problem statement is clear.

You have production scars, not just prototypes. Whatever you've built — browser automation, voice pipelines, orchestration systems — real users hit it, edge cases surfaced, and you fixed them.

You've gone deep in at least one of the four areas. Browser agents, voice agents, multi-agent orchestration, or memory systems. Depth in one matters more than surface coverage across all four.

You've shipped something from zero. A side project that got traction, a new product inside a large company, a founding role — you know what it feels like to own something from nothing and see it through.

Your spike is visible. Tier-one school or company, ex-founder, published research, hackathon wins, exceptional academic record — something in your profile or your conversation signals that you go deep and stay there.

You've stayed long enough to see things break. You don't jump ships every twelve months. You've been around long enough to hit the hard problems that only appear after the first version ships.

You work independently once the problem is clear. High-level direction is enough. You scope, build, and ship without needing someone to check in at every step.

You move with urgency by default. Not because someone is watching, but because you understand that speed is the variable that compounds. The team operates on internal urgency, not external pressure.

You're genuinely curious about what you build. Side projects, blogs, active engagement with what's happening in the field — you build and learn for fun, not just for work.

Ship to production, not to staging. The bar is real users on real systems. Demos and prototypes are not the output.

Pick it up and move. Once the problem statement is clear, you own it. The team does not manage your work; you manage your work.

Plan on Monday, review on Saturday. Two structured touchpoints anchor the week. Everything in between is yours to run.

Six days, high intensity. This is not a company where you ease in over a few months. The expectation is urgency from day one, and it stays.

Collaborate with the right people, not all the people. You pull in whoever the problem needs. You don't wait for a meeting to be scheduled.

Internal urgency is the default. There is no external deadline forcing the pace. The team sets its own bar and holds it.

₹25L–₹60L cash, scaling with experience, plus 0.2% ESOPs vesting over four years. A founding engineering seat on a three-person team, with full ownership of the agent stack from day one. The equity is real; the seat is one of a kind.