Forward Deployed Engineer

Hey - I'm Rishabh, co-founder & COO at Flent. I lead tech, platform, and cash flow here.

If you've rented in an Indian metro, you know the drill. You open 40 tabs, run 12 WhatsApp threads, take 3 broker calls, and still move in with doubt. You never know what's real. The photos, the pricing, the deposit math, the landlord's temperament.

What's broken isn't discovery. It's confidence. At Flent, we're productising every layer of the renting decision until it feels as predictable as any modern consumer experience.

Sometimes, trust also breaks after move-in. Delays, miscommunication, refund edge cases, vendor misses, escalations that repeat because systems don't remember. Every "let me check and get back" is a system failure wearing a human face.

This role exists to kill that class of problem with software.

We're also not building software the old way. We're rewiring our stack around AI—not as a feature, but as infrastructure. Every engineer here will ship at a rate that wasn't possible two years ago. I'm looking for people who want to figure that out with me.

You'll sit where the truth lives: ops, CX, supply, on-ground execution. You'll see what breaks first, what repeats most, and what costs trust. Then you'll build the tools, automations, and platform fixes that permanently eliminate those failures.

Most engineers optimise for clean inputs and well-scoped problems. This role is the opposite. You'll work with messy reality - WhatsApp threads, vendor no-shows, payment mismatches, manually tracked SLAs - and turn that into software that makes the mess go away. If you need a product spec before you can start, this isn't the role.

• The reality → product loop You'll watch where confidence breaks in practice. In ops queues, CX transcripts, vendor SLA misses and translate those into crisp problem definitions, and ship fixes into product and platform. Not tickets. Deployed fixes.

• Internal tools that replace manual work Tools and automations that cut copy-paste, reduce errors, and compress cycle times. If an ops team member is stitching data across three tabs, that's your bug.

• Business observability Unified views connecting billing, ops, and product signals - so we can see what's happening, what it's costing, and where trust is leaking. Dashboards that drive decisions, not decorate standups.

• AI automation with accountability AI where it's safe and high-leverage: ticket routing, summarisation, follow-up generation, and reconciliation. Always with audit trails, permission boundaries, confidence thresholds, and human override. Automation without accountability is faster chaos.

• Systemic fixes for recurring failures If something breaks twice, it shouldn't need heroics the third time. Escalation rules, auto-routing logic, retry policies, SLA enforcement: turning institutional memory into platform behaviour.

• Build internal frontends that ops and CX teams live in (React/Next.js) - triage queues, workflow builders, escalation views, real-time status boards that replace Slack threads and spreadsheets

• Build backend services and integrations (Node/TS or Python), stitching messy inputs- vendor APIs, payment gateways, WhatsApp webhooks, CRM events into clean system behaviour

• Design data models and event pipelines (Postgres + event streams) that make operations measurable, auditable, and debuggable and not just logged, but queryable

• Build business observability with dashboards for SLA adherence, repeat-issue rate, escalation frequency, refund cycle time, vendor turnaround with alerting on drift

• Ship AI automation for ops like intelligent routing, conversation summarisation, follow-up triggers, reconciliation helpers each with audit logs, override paths, and confidence gating

• Use Claude Code daily to learn unfamiliar systems, trace bugs across boundaries, scaffold integrations, and ship fixes without waiting for tribal knowledge

• Go on-ground when needed to visit properties, shadow ops calls, read CX transcripts - to see failure modes dashboards can't capture.

• Has minimum 3–5 years of proven experience in shipping production software where the inputs are messy. Edge cases weren't edge cases but main paths. You've built systems that had to survive real-world chaos, not just pass tests.

• Has strong frontend and backend instincts. You can build the UI that ops lives in and the services behind it and you know which layer a bug is in before opening logs.

• Debugs like it's a sport. Logs, traces, DB state, a Slack screenshot, and a timestamp you've root-caused production issues from less.

• Sees state machines in messy processes. You can look at a human-driven workflow and see where it should be deterministic, where it should be automated, and where it should just be deleted.

• Enjoys AI to work, not to experiment. Claude Code or equivalent is daily infrastructure for learning systems, filling gaps, writing tests, and shipping fixes. Not a side project.

• Learns faster than the problem shifts. New system, new domain, new vendor API with no docs - you close the gap fast. This matters more to us than what you already know.

Educational background doesn’t matter to us. What does is your attitude, your ability to take ownership, and your willingness to do what it takes to make landlords feel taken care of.

Apply now if you feel you are ready for your next challenge!