Srichandan Kota

About

Backend engineer who built and owned an AI-powered financial research product end-to-end, using a typed NestJS/GraphQL backend with LangGraph-style agent routing to produce sourced, structured financial analysis. Emphasizes finance-grade correctness (Zod validation, metric registries, unit/empty-result guardrails) while keeping latency low via batching, caching, and fast token streaming, and has led incremental migrations using strangler/feature-flag/shadow traffic patterns.

Experience

Education

Key Strengths

• End-to-end ownership of an AI-powered financial research backend
• Designs for correctness/guardrails in high-stakes finance (schema validation, metric whitelisting, structured outputs)
• Low-latency API design (GraphQL batching, caching, fewer round trips, fast token streaming)
• Scalable agent/workflow architecture with typed tool calls and sub-agent routing
• Safe incremental migrations (strangler pattern, feature flags, shadow mode)
• Strong API reliability/security practices (timeouts/retries, JWT, RBAC, rate limiting, idempotency, queues)
• Identifies and mitigates silent failure modes (unit mismatches, empty-result summarization) via registries and normalization
• Built and shipped production AI-powered financial research backend end-to-end (API + data + orchestration)
• Strong reliability focus: prevents confident-wrong LLM outputs via registry verification and schema validation
• Designs agent systems with strict planning (JSON) vs execution separation; tools never directly controlled by the model
• Robust error handling with explicit success/retryable/hard-failure states and predefined fallbacks
• Operational rigor: run-level audit trails, step traces, replayability, and monitoring via validation/fallback rates
• Full-stack ownership of an agentic visual workflow builder (UI serialization to JSON graph + backend compilation/execution)
• Solid testing discipline for AI workflows (unit/integration/golden-path) using pytest and deterministic mocks
• Relational modeling for versioned workflows and runs; performance tuning via compound indexes and join reduction