Full-Stack Developer
Own the routing service end-to-end. Build the platform that sends every high-stakes AI decision to the right level of judgment — fast, reliably, at scale.
About the role
Tasq is the orchestrated judgment layer between enterprise AI models and the decisions they can’t afford to get wrong. We route every high-stakes problem — fraud detection, content moderation, model validation — to the right level of human expertise in real time.
You’ll own our core routing service end-to-end. That means the backend systems that deconstruct problems into micro-decisions, the APIs that connect us to a global expert network, and the infrastructure that keeps it all running in production at SLA. You’ll work directly with the team and with customers — no org-chart friction between you and the decision.
Stack: Python, TypeScript, PostgreSQL. ~50% backend, 50% systems. This is a high-ownership, high-autonomy role on a small team with real stakes.
What you'll work on
- Core routing infrastructure — the service that sends every decision to the right human at the right moment
- APIs that connect our platform to a global network of 25K+ credentialed domain experts and 100M+ contributors
- Production monitoring and SLA reliability — these systems run in customer-facing production, not staging
- Data pipelines that feed validation results back into model training
- Internal tooling that lets our ops and CS teams manage complex engagements at scale
What we're looking for
- Strong across the full stack — Python backend, TypeScript frontend, comfortable in both
- Track record of owning complex, end-to-end features from spec to production
- Systems thinking — you reason about reliability, latency, and failure modes, not just features
- Product sensibility — you push back when something doesn’t feel right
- Direct communicator — short written updates, clear opinions, no theater
- Experience with AI/ML pipelines, data ops, or human-in-the-loop systems is a strong plus
What to expect in the interview
We don’t do whiteboard algorithms. We talk through your past work — what you built, why you made the decisions you made, what broke, and what you’d do differently. Then we look at real problems together.
Stages: Intro call (30 min) → Technical conversation with the engineering team (60 min) → Paid work trial (3 days, compensated) → Offer or honest no within a week.
We discuss comp on the first call. No surprises.
Ready to apply?
It takes about 5 minutes. We read every application and respond either way.