Who Shouldn't Apply
- Weak development or backend fundamentals.
- Unable to review, critique, or reject AI-generated code.
- Prompt-only builders without evals, tests, or CI.
- Large unreviewed commits; no typing, docs, or discipline.
What You'll Do
- Own meaningful modules and ship directly with the founders.
- Build AI apps and agents end-to-end (spec > production).
- Experiment with modern AI tools and frameworks and ship fast.
- Make agents reliable through evals, logging, tracing, and controls.
---
**Must-Have**
- Structured problem-solving; clean, typed Python / Go / TypeScript.
- At least one serious AI project shipped (no vibe coding / low-code / n8n).
- Strong Python and/or Go; FastAPI or similar frameworks.
- Hands-on experience with RAG and context engineering.
- Automated or manual evals with regression gates.
- Git discipline; Docker; Linux; CI (GitHub Actions).
- Sound judgment reviewing AI-generated code.
---
**Nice-to-Have**
- System design instincts (latency, cost, throughput tradeoffs).
- Databases: SQL / NoSQL / Redis; schema design and query tuning.
- Observability: OpenTelemetry, feature flags, rate limits, caching.
---
**Perks**
- Full-time offer for strong performers.
- High-ownership environment with a steep learning curve.
- Work alongside a 10x-minded, execution-driven team.
**Selection Process**
Resume / GitHub > Telephonic round > Two technical rounds > **Paid mini-collaboration** (INR 5,000, remote, async; clear brief + rubric; 1–2 days) > Offer.