SUMMARY
Machine Learning Engineer passionate about building ethical and impactful AI solutions.
Expertise spans the full ML lifecycle: from data acquisition and preprocessing, model development,
and rigorous evaluation, to MLOps practices for robust deployment. Specializing in Natural Language
Processing (NLP), Retrieval-Augmented Generation (RAG), predictive modeling, and computer vision.
PROJECTS
Receipt Classifier
PyTorch, EfficientNet-B0, BERT, TrOCR
- Developed a machine learning system that classifies receipts as either real or AI-generated
- Used a hybrid image-text approach combining EfficientNet-B0 for image features and BERT for text features
- Extracted text from images using TrOCR for comprehensive analysis
- GitHub: github.com/Yuguda999/receipt_classifier
RAG PDF Parser
OpenAI, LangChain, Vector Embeddings, Concurrent Processing
- Created a tool that processes PDF documents using Retrieval Augmented Generation (RAG)
- Implemented extraction, analysis, and querying of content using OpenAI's GPT-4 and embeddings models
- Developed concurrent processing, content analysis, and vector embeddings for semantic search
- GitHub: github.com/Yuguda999/rag_pdf_parser
Hybrid LSTM-LR Forecasting
PyTorch, LSTM, Linear Regression, Time Series Analysis
- Implemented a hybrid forecasting model for stock price prediction
- Combined LSTM neural networks and Linear Regression for improved accuracy
- Analyzed historical stock data to predict future price movements
- GitHub: github.com/Yuguda999/hybrid_lstm_lr_forecasting
CrewAI Trip Planner
CrewAI, LangChain, Multi-Agent Systems
- Built an AI-powered trip planning system using CrewAI
- Designed multiple specialized agents with defined roles, backstories, and goals
- Implemented collaborative planning tasks between AI agents
- GitHub: github.com/Yuguda999/crewai_trip_planner
Content Repurposer for Social Media
FastAPI, OpenAI, DALL-E, Celery, Redis
- Developed a production-grade service for AI-powered content repurposing
- Implemented ingestion of raw blog content and transformation into various social media formats
- Utilized OpenAI's GPT models and DALL-E for intelligent content transformation
- Created a scalable architecture with FastAPI, Celery, and Redis
- GitHub: github.com/Yuguda999/content_repurposer
SPAM Classification
Scikit-learn, NLP, Text Classification
- Built a machine learning model to classify emails and messages as spam or legitimate
- Implemented text preprocessing, feature extraction, and multiple classification algorithms
- Evaluated model performance using precision, recall, and F1-score metrics
- GitHub: github.com/Yuguda999/SPAM-CLASSIFICATION