Pune Division, Maharashtra, India
I am a Machine Learning engineer with hands-on experience building AI systems across Computer Vision, Natural Language Processing, and Time Series Forecasting. I recently completed a research internship at IIT Bombay where I worked on document processing and synthetic financial data generation pipelines. My work involved extracting structured information from financial statements and medical reports using OCR, PDF parsing tools, and pattern matching techniques. During the internship I designed data processing workflows and experimented with statistical approaches to simulate realistic financial transaction data. I also built a Streamlit-based prototype application for document upload, transaction extraction, and structured data visualization. I enjoy building complete machine learning pipelines including: • Data preprocessing • Feature engineering • Model training and evaluation • Model comparison and analysis Projects I have worked on include: • Pneumonia Detection using CNN and Transfer Learning • AI-Driven Intrusion Detection System using ML models • Tweet Sentiment Analysis using NLP pipelines • Aircraft Miles Forecasting using ARIMA and SARIMA models Currently I am strengthening my foundations in Machine Learning while exploring Explainable AI, Deep learning and genai. Technical Interests: Machine Learning | Deep Learning | NLP | Computer Vision | Explainable AI | GenAI | Agentic AI. Open to opportunities in: Machine Learning Engineering | AI Engineering | Data Science 📫 Email: [[email protected]](mailto:[email protected]) 💻 GitHub: github.com/AjinkyaKhalikar
Worked on document processing and synthetic financial data generation pipelines for structured data extraction and modeling. • Built document processing pipelines to extract structured information from credit card statements and medical reports using OCR, regex pattern matching, and PDF parsing tools (Tabula, PDFPlumber, Tika). • Designed MongoDB database schema to store demographic patterns, transaction behavior, and income-based spending distributions. • Developed a prototype synthetic financial transaction generator combining demographic attributes with statistical spending patterns. • Built a Streamlit-based web application for document upload, transaction extraction, and categorized data visualization. • Implemented secure file handling, encryption, and PostgreSQL integration for storing processed data. • Conducted experimentation with statistical distributions to simulate realistic financial transactions.