Vetted Software Developers in Washington

“Built and deployed a production chatbot on osvaldocalles.com and iterated through real-world LLM engineering issues: model quota/cost tradeoffs (migrating to Nova Pro), RAG accuracy via semantic chunking, AWS IAM/guardrail/security pitfalls, and Lambda/API Gateway streaming constraints (prefers JS for streaming layer). Experienced with agent orchestration using Strands SDK (AWS-focused) and LangGraph (Vercel/container deployments), plus evaluation pipelines using LLM-as-evaluator, dashboards, and staged model rollouts.”

“Full-stack engineer (Snowflake) who shipped an AI/LLM-powered data exploration product end-to-end, spanning Spring Boot/Python services and a polished React UI with streaming responses and robust fallbacks. Experienced operating high-scale AWS deployments (Docker/Kubernetes, SNS/SQS, RDS Postgres, CloudWatch, Jenkins CI/CD) supporting thousands to tens of thousands of concurrent users, including handling real traffic-spike scaling incidents.”

“Backend engineer who built Flask-based internal APIs supporting GenAI-driven provisioning/diagnostics (Outpost/AWS Outposts-like environment), with deep hands-on optimization across Postgres/SQLAlchemy (2s to <200ms endpoint improvement). Experienced integrating ML/LLM workflows via AWS SageMaker and Bedrock, and designing multi-tenant isolation plus high-throughput Redis-backed background task pipelines (minutes to seconds).”

“Engineering leader with experience defining solutions from business requirements through detailed specifications and implementation, emphasizing cost-aware technology selection. Has led architectural changes including adding IBM Cloud alongside AWS for budget reasons and integrating caching/messaging to improve availability and performance, and describes scaling distributed teams via experienced DevOps/QA hires and structured evaluation.”

“Full-stack engineer who built and owned an AI-assisted job-matching dashboard in Next.js App Router/TypeScript, keeping LLM logic server-side and improving performance via deduplication, caching/revalidation, and streaming (35% fewer duplicate LLM calls; 40% faster first render). Also has strong data/backend chops: designed Postgres models and optimized queries at million-record scale (1.8s to 120ms) and built durable AWS multi-region telemetry workflows with idempotency, retries, and monitoring.”