Senior/Staff Software Engineer

Senior/Staff Software Engineer - AI‑AcceleratedSalary: 40 LPALocation: Hybrid(Chennai)Belief: We believe modern AI tools exponentially compound the output of the very best engineers. Our bar is intentionally high: we hire exceptional builders and give them an AI multiplier.The RoleOwn end‑to‑end delivery of production systems in an AI‑first environment. You’ll turn fuzzy problems into tight specs, orchestrate AI coding tools as collaborators, and ship reliable, secure, observable software that users love.What You’ll DoFrame problems & write crisp specs with clear acceptance criteria, non‑goals, and performance/cost/privacy constraints.Use AI coding tools as force multipliers (generate, refactor, review) while you own architecture, correctness, and maintainability.Engineer verification: property‑based tests, mutation testing, golden tests, and CI gates that prevent regressions.Design robust architectures (modular boundaries, contracts, versioning, idempotency, retries/backoff, feature flags, canaries).Build data & knowledge foundations: schemas, data contracts, quality checks; pragmatic RAG/embeddings where appropriate.Operate AI features: tracing, offline/online evals, drift monitoring, guardrails, human‑in‑the‑loop feedback.Harden security & safety: threat models for AI chains, least‑privilege tool access, I/O validation, SBOM/supply chain hygiene.Optimize performance & cost: profiling, caching/batching/streaming, p95/p99 targets, cost budgets, graceful degradation.Instrument reliability: SLIs/SLOs, structured logging, distributed tracing; circuit breakers and fallbacks.Collaborate crisply: write concise RFCs/ADRs, make reversible decisions, and mentor others in AI‑accelerated workflows.You’ll Thrive Here If You HaveProduct sense & taste: you choose the simple thing that scales; your naming, APIs, and defaults reduce cognitive load.Rapid domain learning: you become credible in unfamiliar domains in days, not months - through targeted research and fast prototypes.Spec‑first, test‑first instincts: you treat specs, tests, prompts, and datasets as versioned artifacts.Systems thinking: you design for failure, latency, and cost from the start.Clean fundamentals: algorithms/DS, concurrency, networking/OS, DB internals, type systems - so you can audit AI output.Core Skills We Care AboutProblem Framing & SpecificationContext Engineering & Workflow Design (long term memory, HITL, tool use)Verification Engineering (property‑based, metamorphic, mutation)Architecture & Integration (contracts, compatibility, migration paths)Data/Knowledge Engineering (schemas, quality, governance; pragmatic RAG)LLMOps & Evaluation (telemetry, eval sets, drift, feedback loops)Security, Privacy & AI‑Specific Threat Modeling (prompt injection, data exfiltration, tool sandboxing)Performance & Cost Engineering (profiling, caching, batching, streaming)Reliability & Observability ( tracing, dashboards, incident response)Uncertainty‑Aware UX (confidence cues, sources, safe fallbacks)Communication & Decision‑MakingNice to HaveEntrepreneurial experience (founded a company or led a 0→1 product).Prior ownership of an AI‑infused production system with evals/guardrails/telemetry.How We WorkAI‑first, spec‑first, test‑first. Code is cheap; correctness and systems are not.Small, senior teams. High ownership, fast feedback, low ceremony, strong documentation.Measure everything. If we can’t observe it, we don’t ship it.QualificationsSenior/Staff‑level impact demonstrated through shipped, operating software.Evidence of product judgment and taste in API/UX design.Track record of learning new domains fast and making good decisions under ambiguity.Comfortable supervising AI coding tools and holding the line on quality, security, and reliability.Equal OpportunityWe value diverse perspectives and are an equal‑opportunity employer. If you’re exceptional but don’t tick every box, please apply.How to Apply - Project Submission OnlyComplete the take-home project in https://tinyurl.com/yjepr8xcSubmit your GitHub repo along with your CV. Use this Google form https://tinyurl.com/2vew6ncf