Software Developer & AI Engineer
I build enterprise-grade CRMs and AI pipelines. From integrating REST APIs across 100,000+ records to fine-tuning multilingual LLMs on HPC clusters, I focus on scalable systems and data-driven engineering.
“I don’t just write code, I think through problems.” My approach combines algorithmic rigor with a product-driven mindset. Whether it’s solving 300+ LeetCode challenges or deploying full-stack architectures, I strive for code that is as elegant as it is efficient.
Career Path
Work
A selection of my architectural and engineering work across different domains.
End-to-end evaluation and fine-tuning pipeline for 1K+ BLEnD multilingual QA samples using Qwen. Applied SEM-B/SEM-W scoring and distributed GPU experimentation on OSCER HPC. Analyzed reasoning consistency across 200+ prompt variations and 20+ training runs.
Trained an NVIDIA JetBot to autonomously follow a predefined path using real-time visual input. Collected and preprocessed 10,000+ images; implemented and compared Custom CNN, MobileNet, and ResNet18. Custom CNN achieved best performance in accuracy, smoothness, and responsiveness.
Capabilities
I specialize in building scalable backends and intelligent frontends, leveraging modern tools to ensure reliability and performance.
Exploring LLM fine-tuning and retrieval-augmented generation (RAG) architectures.
Transitioning AI workloads for serverless scaling and data-heavy performance.
Education
I’m always open to discussing technical challenges, innovative projects, or research collaborations.
bhagyeshparmar50@gmail.com