Manasa Pantra

About

Product-focused full-stack engineer at startup (Zippy) who shipped a production multi-agent AI system for restaurant operations plus payments workflows. Built end-to-end: RAG grounded on a Notion knowledge base, structured function-calling task routing, FastAPI/JWT multi-tenant backend, and a polished React+TypeScript owner dashboard. Has real production incident experience (duplicate Stripe webhooks) and reports ~94% task-routing accuracy under load.

Experience

Education

Key Strengths

• Shipped multi-agent AI system to production for restaurant operations
• Improved production AI reliability via RAG grounding + clarification flows to reduce hallucinations
• Designed robust task routing using structured function calling with strict schemas (valid JSON before actions)
• Built customer-facing AI menu management with validation + confirmation layers for high-impact changes
• Owned FastAPI backend with JWT auth, multi-tenant authorization, versioned APIs, and standardized errors
• Production reliability engineering: idempotent Stripe webhooks, retries, async non-blocking external calls
• Performance/scale practices: caching, horizontal scaling behind load balancer, monitoring latency/error rates
• React+TypeScript dashboard with typed hooks, consistent loading/error UX, optimistic updates with rollback
• PostgreSQL modeling for conversations/intents/knowledge-graph entities with indexing and query-driven tuning
• Handled real production incident: duplicate Stripe webhook events under high traffic; fixed via event-id checks and faster webhook responses
• Reported ~94% accurate task routing under production load and continuously monitored/improved it
• Built and deployed an LLM assistant that can take real operational actions (not just chat)
• Strong focus on production reliability; implemented retrieval-only answering to reduce hallucinations
• Designed guardrails and verification steps before making business-critical updates
• Effective multi-agent orchestration for decomposing complex requests into parallel tasks with all-or-nothing updates
• Practical testing strategy covering edge/failure scenarios plus agent-level isolation testing
• Metrics- and logging-driven evaluation in production (accuracy, success rate, response time; decision traceability)
• Pragmatic model/retrieval/prompt selection based on accuracy/speed/cost/reliability tradeoffs
• Able to translate non-technical stakeholder needs into usable product features via demos and iterative feedback

Reference Highlights

• Strong at understanding requirements
• Consistently delivers production-ready, customer-facing features
• Significant impact on team despite joining later
• Built internal tools used by the entire team
• Reliable and consistent; never missed a deadline
• Proactive progress communication
• Improved operational efficiency via active-orders monitoring tool
• Helped ensure strong customer experience by catching order/delivery issues quickly
• Strong end-to-end full-stack execution
• Writes good modular code
• Rigorous testing; code is dependable
• Exceptional React/TypeScript frontend results
• Professional, intuitive design/UX skills
• Takes feedback well and iterates effectively
• Strong PostgreSQL/SQL and relational modeling
• Takes initiative in AWS deployment/operations
• Performs well during production issues/incidents
• Drives ideas from 0 to 1 in ambiguous environments
• Collaborates closely with founders/leadership with minimal oversight