Open to opportunities

Hemish Radadiya

@hemishradadiya

Message

I build production GenAI and full-stack microservices to automate growth and personalize experiences.

United States

Message

What I'm looking for

I’m looking for a team where I can ship production GenAI systems and full-stack microservices—automating workflows, improving personalization, and mentoring through thoughtful engineering in a fast-moving, measurable environment.

I’m a Full Stack AI Engineer focused on shipping production GenAI systems. At Sarvay, I architected end-to-end LangChain/RAG/LLM pipelines for marketing automation, eliminating manual campaign effort by 40% while improving lead targeting across multi-channel workflows.

I also turn research into usable products. As a Graduate Research & Teaching Assistant at Lamar University, I trained CNNs (CNNs, YOLO, U-Net) on aerial disaster imagery (92% mAP, 0.89 F1-score), deployed ML models as REST APIs with FastAPI, and built LLM-powered RAG apps using LangChain and Azure OpenAI.

Before AI-focused work, I built reliable enterprise software. At Casepoint LLC, I scaled distributed SaaS to 10,000+ concurrent users, integrated Redis caching and Elasticsearch to cut API response time by 45%, and delivered 150+ production issues within a 24h SLA while mentoring engineers through 40+ code reviews.

Experience

Work history, roles, and key accomplishments

Current

Full Stack AI Engineer

Current

Sarvay

May 2025 - Present (1 year 1 month)

Architected end-to-end GenAI pipelines with LangChain, RAG, and LLMs for production marketing automation, eliminating manual campaign effort by 40% and improving lead targeting. Built FastAPI + Azure OpenAI microservices with agentic AI workflows and deployed with Docker/Kubernetes and zero-downtime CI/CD.

Lang Chain RAG Azure OpenAI fastAPI Agentic AI Microservices Docker Kubernetes CI CD

Graduate Research & Teaching Assistant

Lamar University

Jun 2025 - May 2026 (11 months)

Trained CNNs, YOLO, and U-Net in PyTorch for aerial disaster imagery, achieving 92% mAP and 0.89 F1 for flood and structural damage detection. Reduced training time from 72h to 18h using multi-GPU and FP16 optimization, and deployed models as REST APIs with FastAPI/Docker while teaching 80+ students.

PyTorch CNN YOLO U Net Multi GPU Training FP16 Optimization fastAPI Docker RAG

Software Engineer

Casepoint LLC

Apr 2023 - Jul 2024 (1 year 3 months)

Scaled an enterprise legal SaaS to support 10,000+ concurrent users with 99.5% system uptime using C# and ASP.NET MVC. Improved performance by 45% response time via Redis caching and Elasticsearch over 10M+ documents, and resolved 150+ production issues within 24h SLA while mentoring 5 engineers.

C ASP.NET MVC Vue.Js PostgreSQL Redis Elasticsearch Microservices SQL Optimization Performance Tuning CI CD

Associate Software Engineer

Casepoint LLC

Jun 2021 - Apr 2023 (1 year 10 months)

Delivered 50+ bug fixes and enhancements across C#, ASP.NET, and SQL Server in Agile sprints. Migrated legacy UI to a Vue.js SPA, improving page load by 35%, and resolved 80+ tickets while maintaining 99.5% uptime.