Akarshika Tripathi

About

Python backend developer from Larix Technologies who built and scaled microservice APIs for an omnichannel messaging SaaS (WhatsApp/Instagram/Facebook) and led production performance fixes during peak traffic, cutting webhook latency ~50%. Also shipped applied AI products end-to-end: a RAG-based PDF assistant (LangChain + Mixtral via Groq + React) and a BI agent that plans/executes/verifies multi-step analytics with strong guardrails and auditability.

Experience

Key Strengths

• Debugged and resolved peak-traffic reliability/performance incident in event-driven messaging system (DB + async pipeline bottlenecks)
• Database performance tuning at scale (indexes, batching ORM calls, optimized joins, reduced columns fetched)
• Async/background architecture improvements (split Celery tasks, idempotency on retries, moved slow third-party calls off request path)
• Measured production impact: ~50% webhook latency reduction and ~35% message processing lag reduction
• End-to-end ownership of applied AI product (DocuGenie) across backend RAG pipeline and React UI
• Designed reliable agent workflow with explicit plan/execute/verify phases and deterministic computation for BI questions
• Strong production guardrails for AI systems (observability, evaluation sets, grounding thresholds, fallbacks, auditability)