AI/ML Engineer & Full-Stack Developer | Computer Science @ University of Pittsburgh
I'm passionate about building production-ready LLMs, AI-powered applications, and modern web solutions. Currently developing a fine-tuned business planning model that has helped 100+ startup founders accelerate their business development.
- π§ Personal Knowledge Assistant MCP Server: A comprehensive Model Context Protocol server that transforms personal information management across email, social media, documents, and productivity metrics
- π€ Building custom LLM implementations with transformer architectures and RAG systems
- π¨ Creating immersive web experiences with React, Next.js, and Three.js
- π Open-source educational resources for LLM development
April 2025 - Present
- Built GPT-style transformer from scratch using PyTorch with custom tokenizer and attention mechanisms
- Fine-tuned models on instruction datasets, deployed via Chainlit for real-time interaction
- Created open-source educational resources with LLaMA-based scoring
Freelance | August 2024 - March 2025
- Developing AI solutions for SEO & lead generation with RAG chatbots using OpenAI API
- Building context-aware customer support bots with enhanced website navigation
2024 - Present
- Building modern web apps with React/Next.js, TypeScript, and Tailwind CSS
- Creating 3D visualizations with React Three Fiber and particle effects
A comprehensive Model Context Protocol server that revolutionizes personal information management
- π§ Email Intelligence: Smart email management with AI-powered analysis and composition
- π Social Media Integration: Multi-platform support (Twitter, LinkedIn, Facebook, Instagram) with performance analytics
- π Personal Analytics: Track productivity metrics, habits, and goals with data-driven insights
- π Privacy-First: End-to-end encryption, local processing, GDPR compliant
- π οΈ 7 Powerful MCP Tools: Including email analysis, social media management, project tracking, and insight generation
- π§ AI-Powered Insights: Behavioral pattern detection, predictive analytics, and relationship mapping
Fine-tuned Mistral-7B model designed to guide startup founders with AI tool recommendations
- π₯ Helped 100+ startup founders accelerate business development
- π Trained on ~1,500 GPT-4-generated prompt-response pairs
- β‘ Implemented QLoRA (4-bit) and Flash Attention 2 for efficient training
- π― Provides structured roadmaps, clarifying questions, and contextual AI tool recommendations
Modern NextJS website showcasing AI capabilities through impressive 3D visuals
- π¨ Built with Next.js 14+, TypeScript, and Tailwind CSS
- π 3D graphics using React Three Fiber with custom particle systems
- π± Optimized for mobile with lazy loading and comprehensive SEO
- π€ Features MCP servers with agentic AI systems and Zod form validation
End-to-end implementation of a Large Language Model using Python and Jupyter Lab
- ποΈ Built transformer architecture from scratch with multi-head attention
- π§ Implemented custom tokenizer and training pipeline
- π Created comprehensive documentation for educational purposes
Expertise: Fine-tuning (QLoRA, LoRA), RAG Systems, Transformer Architecture, Flash Attention, Model Evaluation, Prompt Engineering, Model Context Protocol (MCP), Chainlit, Axolotl
- π University of Pittsburgh - B.S. Computer Science (2025-2029)
- π Harvard CS50 - Introduction to Computer Science (Completed)
- π Northern Virginia Community College - Web Design and Development
- π Certifications:
- Certificate of Completion - Cybersecurity: Protecting Our Digital Future
- Certified Entry-Level Web Developer
π‘ Currently exploring: Advanced MCP server implementations for personal knowledge management, agentic AI systems, and privacy-preserving ML techniques
π Fun fact: I built my first transformer model from scratch at 18!