Developed “Lovelace,” an autonomous cryptocurrency trading bot that uses Deep Reinforcement Learning to execute trades on the Binance exchange. Available at https://www.twine.net/signin

• Engineered an automated Bitcoin trading bot using a Deep Reinforcement Learning (Actor-Critic) model to optimize trading strategies.

• Integrated multiple technical analysis indicators (SMA, EMA, RSI, MACD, Bollinger Bands) to inform the agent’s decision-making process.

• Built the model with a shared Convolutional Neural Network (CNN) to extract features from a sliding window of market data, including OHLC prices and portfolio state.

• Connected the agent to the Binance API for real-time data ingestion and automated order execution in a live production environment.

• Developed a simulated environment for robust backtesting, consistently outperforming a “hold” strategy in tested scenarios.

Skills: Deep Reinforcement Learning · REST APIs · Convolutional Neural Networks (CNN) · Python Development · TensorFlow

• Engineered a C# WPF application to control the volume of individual Windows applications using an external hardware controller.

• Developed a robust serial communication protocol to interface with an Atmega microcontroller, enabling real-time, tactile audio adjustments.

• Implemented an application selector and a customizable RGB lighting system to provide visual feedback and a personalized user experience.

• Built the application using .NET Core, creating a modern and responsive user interface for managing audio settings.

• Successfully integrated software and hardware to create a seamless and intuitive solution for advanced audio control.

Skills: C# · .NET Core · WPF Development · Arduino

• Engineered an AI-powered agent that analyzes user documents (PDF, TXT, CSV) to provide conversational, context-aware answers.

• Built the core Retrieval-Augmented Generation (RAG) pipeline using LangGraph to create a sophisticated, stateful agent.

• Integrated support for both powerful cloud models (Google Gemini) and local LLMs (Ollama) for flexible and scalable performance.

• Implemented a standout feature to automatically generate PowerPoint presentations from the conversational context using the Model Context Protocol (MCP).

• Utilized ChromaDB to create an efficient vector store for document indexing and retrieval, and developed a Flask-based web UI for a seamless user experience.

• Available at: https://www.twine.net/signin

Skills: Langgraph · Generative AI · Python Development · Flask · ChromaDB · LLM Finetuning

Developed a sophisticated multi-agent AI system designed to streamline the clinical intake process and provide practitioners with data-driven diagnostic hypotheses and treatment plans.

Architecture: Engineered a coordinated multi-agent pipeline using LangChain and LangGraph, featuring specialized agents for patient anamnesis, clinical evaluation, and therapeutic research.

Knowledge Integration: Implemented Retrieval-Augmented Generation (RAG) using ChromaDB to index the Medline database, coupled with a real-time PubMed web-crawling agent to retrieve state-of-the-art treatment guidelines.

Performance & Validation: Rigorously evaluated the system against a dataset of 250 simulated patient cases, achieving a 90.1% total diagnosis accuracy and a mean clinical quality score of 83.4/100.

Safety & Compliance: Designed an “AI Patient” simulation framework for testing and integrated emergency detection protocols to identify critical symptoms (e.g., chest pain, slurred speech) during the anamnesis phase.

Tech Stack: Python, Google Gemini, Ollama (Local LLMs), FastAPI, ChromaDB, LangGraph