Haritha Peddi

About

Backend/AI engineer with production experience in payments/reporting systems and high-scale Node/NestJS services on AWS (ECS/ALB) using PostgreSQL, Redis, Kafka, Prisma, and Datadog. Shipped applied AI features including a Zendesk-embedded support copilot (summarization, draft replies, internal doc retrieval, playbook next steps) and an LLM-driven ops workflow agent with robust error taxonomy, retries/escalation rules, and auditability.

Key Strengths

• Built and operated high-scale backend services (~8–12M requests) on AWS ECS with PostgreSQL/Redis/Kafka
• Deep performance debugging: traced P95 latency spikes to Prisma N+1 queries and missing/incorrect composite index usage; used Datadog APM and EXPLAIN to remediate
• End-to-end ownership of an AI support copilot embedded in Zendesk (ideation, model selection, integrations, deployment, feedback loop)
• Designed reliable multi-step LLM-driven ops automation with explicit planner/executor/verifier separation and deterministic error-handling policies
• Production engineering discipline: risk-based testing, contract testing, failure-path testing, and strong observability/on-call practices
• AI production guardrails: prompt/model versioning, continuous evaluation/monitoring, fallbacks, human-in-the-loop defaults, and audit logs

Experience

Education

Skills

Languages