Back-End Developer

● Cut AI integration complexity by 60% and eliminated user-facing failures during provider outages by designing a provider-agnostic microservices architecture — a routing/gateway service fronting pluggable OpenAI, Gemini, and Anthropic adapters — with automatic failover behind a single abstraction. ● Achieved zero user-facing AI failures during live LLM outages, measured across all production incidents, by designing a three-tier resilience pattern (Retry → RAG Fallback → Rule-Based Floor) and defining it as the team's standard failure-handling architecture. ● Reduced dashboard response times from 12s to under 2s (83% improvement) by owning end-to-end PostgreSQL and MongoDB query and performance optimization, indexing strategy, and MeiliSearch rollout. ● Sustained sub-second API response times under thousands of concurrent async jobs by designing an event-driven processing architecture on BullMQ and Redis and setting the technical direction for high-throughput workloads. ● Shortened release cycles and lifted developer productivity, measured by faster time-to-deploy, by building GitHub Actions CI/CD pipelines with automated test coverage as a standing testing strategy. ● Held zero cross-tenant data exposure incidents across the platform's lifetime by architecting secure REST and GraphQL APIs on OAuth 2.0/JWT and enforcing tenant isolation as a non-negotiable design standard. ● Delivered a complete timesheet module end-to-end, measured by full Jest unit-test coverage at handoff, using GitHub Copilot as the primary implementation tool under the same review and correctness bar as hand-written code. ● Built an internal ATS platform, measured by shipped production features, using Claude Code as an AI pair-programmer for scaffolding and Gemini 2.5 Flash for document parsing and the conversational assistant, while retaining full ownership of architecture and review. ● Raised engineering consistency across the team, measured by review-standard adoption, by mentoring backend engineers through architecture and code review covering both AI-assisted and hand-written code. ● Authored “Designing for AI Failure,” a technical case study documenting the platform's three-tier resilience architecture with system diagrams and a reversible vector-store ADR, used internally to align technical direction and onboard engineers.