Hi, I'm

Jay Ghiya

Builder. Engineer

Currently building unoplat-code-confluence — an open source code intelligence platform for better AI workflows.

About Me

I’m a full-stack engineer currently learning applied machine learning and deep learning through building unoplat-code-confluence, an open-source product. I’m expanding into LLM engineering, working with retrieval-augmented generation (RAG), fine-tuning techniques, and developing autonomous agents that solve complex problems. Technologies & Skills:
  • React & TypeScript
  • Python
  • Design Systems & UX
  • Backend Engineering
  • Docker & Kubernetes
  • CI/CD Pipelines
  • Observability
  • Retrieval-Augmented Generation
  • Fine-tuning Techniques
  • Autonomous Agents
  • Vector Databases
  • Product Engineering
  • Project Management

Featured Projects

Python Microservice Workflow Orchestration Agents RAG Tree-sitter LLM Engineering
Unoplat Code Confluence

A codebase intelligence platform that uses formal grammars and schemas to create reliable, structured context from your code. Unlike brittle heuristics or LLM guesswork, it enables reliable, auditable, and efficient AI workflows built on trustworthy foundational context.

Core Values:

  • Precision: Grammar-based analysis, not regex or LLM guesswork
  • Efficiency: Process large codebases without drowning in noise
  • Reliability: Consistent results you can trust
  • Self-hosting: Your code stays on your infrastructure
  • Transparency: Open Source

First AI Use Case: Auto-updating AGENTS.md per repository that captures sections like Project Structure, Development Workflow, and Business Logic through deterministic code grammar and agents.

View Demo