At a glance
At a glance
Building production AI systems grounded in evaluation metrics, measurable impact, and reliable delivery.
How I build
Rigorous evaluation and testing to ensure models perform as expected before release.
Fine-tuning accuracy
+20%
Domain-specific accuracy with QLoRA fine-tuned Mistral 7B
Click to view project →
Retrieval latency
-15%
Hybrid search optimization with Pinecone and ChromaDB
Click to view project →
AI projects
8
Production-ready AI applications deployed
Click to view project →
Full-stack experience
4+ years
MERN stack and AI-powered web applications
Performance focused
Optimized retrieval and caching strategies keep latency low while preserving factuality.
Safety & reliability
Structured output validation and monitoring to ensure safe, reliable behavior in production.
Evidence over claims
Projects with measurable impact
Each project built around a real problem, with measurable outcomes.
SheetBrain-AI
01/2025 — 02/2025Problem
Enterprise teams needed automated spreadsheet error detection and formula auditing with policy compliance checks.
Solution
Engineered an AI-powered formula auditing tool using TypeScript and Google Apps Script. Designed real-time recommendation engines with policy compliance checks and intelligent formula insights. Created seamless integration between cloud spreadsheets and LLM APIs.
Legal-FAQ-Assistant
10/2025 — 11/2025Problem
Legal Q&A systems lacked domain-specific accuracy and data privacy compliance for sensitive legal queries.
Solution
Fine-tuned Mistral 7B using QLoRA, achieving 20% increase in domain-specific accuracy. Architected a secure pipeline with AES-256 encryption and containerized deployment via Kubernetes. Optimized model inference for accurate legal responses.
Multimodal-AI RAG App
05/2023 — 07/2023Problem
Organizations needed to query multi-format documents with fast retrieval and contextual reasoning across different file types.
Solution
Developed full-stack RAG application with FastAPI and React.js. Implemented hybrid search pipeline using Pinecone and ChromaDB, reducing retrieval latency by 15% through optimized metadata filtering. Integrated Google Gemini API for contextual reasoning.
Babel-Fish-Assistant
08/2023 — 09/2023Problem
Real-time voice translation systems lacked seamless integration of speech recognition, translation, and natural voice synthesis for cross-language communication.
Solution
Developed a production-ready real-time voice translator integrating Deepgram STT and ElevenLabs TTS with GPT-4o-mini for high-accuracy cross-language processing. Built full-stack application with Next.js 14 and Flask backend for low-latency audio streaming.
Multi-Agent-AI-System
10/2023 — 12/2023Problem
Complex AI tasks required autonomous orchestration of multiple specialized agents with tool-based workflows and collaborative decision-making.
Solution
Engineered a Multi-Agent AI System capable of autonomous task decomposition and execution using tool-based workflows and collaborative agent architectures. Implemented agent coordination patterns with TypeScript for reliable multi-step task completion.
AI-Powered-Meeting-Assistant
01/2024 — 02/2024Problem
Manual meeting note-taking consumed significant time while missing key action items and decisions requiring follow-up.
Solution
Built a modern meeting intelligence platform that transcribes live audio and generates actionable summaries using state-of-the-art NLP models. Implemented real-time speech-to-text processing with Python and Flask, reducing manual note-taking time.
AI CRM Automation System
02/2025 — 03/2025Problem
Manual lead triage, enrichment, and follow-up drafting slow down high-velocity sales pipelines.
Solution
Built an AI CRM automation pipeline using n8n for orchestration, Claude 3 Haiku for qualification and scoring, Supabase for persistence, and a realtime Next.js dashboard. Automated lead summaries, email drafts, and next-action suggestions end-to-end.
AI Lead Conversion System for Real Estate
03/2025 — 04/2025Problem
Disconnected CRM tools create leakage between first touch and booked appointment and make pipeline consistency hard to enforce.
Solution
Built a multi-tenant conversion platform that unifies conversational capture, AI qualification, deal lifecycle management, and analytics. Enforced tenant isolation and deterministic stage transitions in Supabase-backed APIs, with optional async delivery via FastAPI and Redis.
Technical credibility
Stack, tools, and how they are used
Interactive honeycomb shows exactly how each technology is applied in production workflows. Click to see implementation.
Large Language Models
Gemini, LLaMA, Mistral - fine-tuning and inference
→ Legal-FAQ-Assistant
LLM Fine-Tuning
LoRA, QLoRA, PEFT for domain optimization
→ Legal-FAQ-Assistant
RAG Architectures
Hybrid search and semantic retrieval
→ Multimodal-AI RAG App
Vector Databases
Pinecone, ChromaDB, FAISS for embeddings
→ Multimodal-AI RAG App
FastAPI
Asynchronous AI inference pipelines
→ Multimodal-AI RAG App
React.js & Next.js
AI-powered full-stack applications
→ Babel-Fish-Assistant
Python
AI systems and backend development
→ AI-Powered-Meeting-Assistant
Docker & Kubernetes
Containerization and AI deployments
→ Legal-FAQ-Assistant
Prompt Engineering
Structured outputs and agentic workflows
→ Multi-Agent-AI-System
PostgreSQL
Database architecture and SQLAlchemy ORM
→ SheetBrain-AI
Experience
Roles, outcomes, and delivery
Where I've applied my skills
Full Stack Developer
Freelance01/2021 — Present- Built an AI document processing pipeline for a legal tech startup, reducing manual review time with automated extraction and classification.
- Developed a RAG-based internal knowledge base for a consulting firm, enabling semantic search across 500+ internal documents.
- Delivered an AI CRM automation system for a real estate agency, automating lead qualification and follow-up workflows end-to-end.
- Architected real-time voice translation system integrating Deepgram STT, GPT-4o-mini, and ElevenLabs TTS for cross-language communication.
- Designed and deployed full-stack web applications using MERN stack (MongoDB, Express.js, React.js, Node.js) with JWT/OAuth authentication.
- Optimized website performance and Core Web Vitals, achieving 90+ Lighthouse scores on production deployments.
- Built multi-tenant SaaS platform with Supabase, enforcing tenant isolation and deterministic deal pipeline workflows.
- Integrated production-grade LLM APIs (OpenAI, Anthropic, Google) for chatbots, document analysis, and automation workflows.
- Managed deployments on Vercel, Netlify, AWS, and Heroku with CI/CD pipelines and monitoring.
Education & Certifications
Academic background and training
IBM-certified AI developer with 4+ years building production AI systems.
Professional Certificate, AI Developer
- Completed an intensive 10-course program focusing on the Full Stack AI Development lifecycle.
- Developed core competencies in Generative AI, Prompt Engineering, and building AI-powered applications.
- Gained hands-on experience in Python for Data Science, FastAPI/Flask, and creating specialized AI chatbots.
- Completed rigorous project-based training in modern web technologies and software architecture.
- Specialized in the MERN stack and professional agile development methodologies.
Bachelor of Science, Physics & Chemistry
- Foundational training in mathematics and physical sciences.