Joel

AI Systems Engineer

Freelancing Career • Part time

Jan 2025 - Mar 2026 • 1 yr 2 mos

Architected and deployed LLM-powered automation systems for 5 enterprise clients, reducing manual document processing time by 62% (from 4 hours to 90 minutes per day). Built contextual AI assistants with knowledge retrieval capabilities using LangChain, OpenAI GPT-4, and pgvector, achieving 94% accuracy on domain-specific queries. Integrated LLM APIs (OpenAI, Anthropic Claude) into production web applications serving 50,000+ monthly users with 200ms average response time. Designed scalable vector search infrastructure using PostgreSQL with pgvector, indexing 1M+ embeddings with sub- 100ms query latency (95th percentile). Developed RAG pipelines for legal and financial document analysis, reducing information retrieval time by 78% (from 45 minutes to 10 minutes per document). Implemented prompt engineering framework with version control and A/B testing, improving response relevance by 35% across 8 different LLM models. Containerized AI services with Docker and deployed on AWS ECS (Fargate), achieving 99.95% uptime and reducing inference costs by 28% through optimized batching. Created evaluation pipelines for LLM outputs (ROUGE, BERTScore, custom rubrics), reducing hallucination rate from 15% to 4% over 6 months.

Senior Software Engineer

Crunchbase • Part time

Jan 2022 - May 2024 • 2 yrs 4 mos

Led development of scalable web applications and backend services for international SaaS and e-commerce clients, serving 2M+ monthly active users across 3 platforms. Engineered full-stack applications using React, Next.js (App Router), Node.js, and TypeScript, achieving 95+ Lighthouse scores for performance and accessibility. Built scalable backend APIs using Node.js and Express, handling 10,000+ concurrent requests with average response time of 85ms (p95). Structured and optimized PostgreSQL database schemas with advanced indexing, partitioning, and query optimization, reducing query latency from 450ms to 95ms (79% improvement). Implemented secure authentication systems using OAuth2 and JWT, processing 500,000+ daily authentication requests with 99.99% uptime. Architected microservices-based architecture (12 services) to improve system modularity and maintainability, reducing deployment conflicts by 85%. Containerized backend services using Docker and deployed through CI/CD pipelines (GitHub Actions), reducing deployment time from 2 hours to 15 minutes (87% reduction). Collaborated with distributed engineering teams across 4 time zones, participating in architecture design reviews and sprint planning for 10+ major releases. Migrated legacy monolithic application to microservices, reducing mean time to recovery (MTTR) from 2 hours to 25 minutes. Implemented comprehensive monitoring with Prometheus and Grafana, reducing incident detection time from 15 minutes to 3 minutes. Reduced cloud infrastructure costs by 32% ($180K annually) through resource optimization and spot instance adoption. Conducted 50+ code reviews and mentored 4 junior engineers, 2 promoted to mid-level within 12 months.