San Francisco Bay Area
Software Engineer with almost 5+ years designing scalable Java/Spring Boot microservices, cloud-native DevOps platforms, and AI/ML-powered backend systems. My work spans three domains: backend system design (Java 17, Spring Boot 3, Kafka, PostgreSQL, GraphQL), DevOps platform engineering (AWS EKS, Terraform, Helm, ArgoCD, CI/CD), and applied AI/ML engineering (LangChain, LangChain4j, RAG pipelines, LLM fine-tuning with LoRA/QLoRA, Hugging Face Transformers). Key results: reduced release cycles by 85%, cut DB query latency by 40%, built RAG systems with 85% factual accuracy, and fine-tuned Mistral-7B to 91% F1 on financial NER. Currently completing MS in Computer Science at the University of Central Missouri (May 2026). Open to backend, platform engineering, or AI/ML engineering roles in the US — feel free to connect.
Architected AI-powered RAG pipeline using LangChain4j + FAISS over 500+ financial PDFs — enabling semantic Q&A with 85% factual accuracy and sub-300ms retrieval latency Fine-tuned Mistral-7B using QLoRA/LoRA on 15K labeled financial documents — achieved 91% F1 on NER (vs 73% baseline), deployed on AWS SageMaker with Resilience4j circuit-breaker fallback Decomposed monolith into 6 domain-driven microservices — reduced cross-service deployment coupling by 70%; 50K+ daily transactions at sub-200ms P99 Architected CI/CD pipelines (Jenkins + GitHub Actions) for 8 microservices — cut release cycle from 3 days to 5 hours (85% improvement) Operated AWS EKS cluster with 200+ pods — HPA absorbed 3× traffic spikes while sustaining 99.9% uptime SLA Provisioned full AWS infra (VPC, EKS, RDS, IAM) via modular Terraform — reduced manual provisioning from 2 days to 4 hours Deployed Prometheus + Grafana observability stack — reduced incident MTTD by 60%; SonarQube SAST caught 150+ critical issues pre-production
Prototyped NLP log classifier using DistilBERT (Hugging Face) — reduced mean incident classification time by 35%, feeding structured labels into ServiceNow workflows Configured Spring Cloud Gateway centralizing JWT auth, rate limiting, and routing across 8 microservices — eliminated per-service auth boilerplate Managed 20+ AWS EC2 production servers — right-sizing, Auto Scaling Groups, and cost-optimization reviews Wrote Bash/Python automation eliminating 8+ hours of weekly manual ops across provisioning and log management Replaced O(n²) nested loops with HashMap O(n) lookups — cut large-dataset processing time by 45%