PROFESSIONAL SUMMARY

Data Scientist | GenAI

Data Scientist (ML & Applied AI) with 3 years of experience delivering production-grade machine learning systems that create measurable business impact. Strong background in Python and SQL, owning the full ML lifecycle from problem framing and experimentation to deployment and monitoring. Hands-on experience with LLMs (RAG, fine-tuning, evaluation), reinforcement learning, and forecasting, combined with solid software engineering practices (Docker, CI/CD, MLflow). Proven ability to work in fast production cycles, collaborate with product and business stakeholders, and translate complex data problems into scalable, reliable solutions.

TECHNICAL SKILLS

Machine Learning & AI:

Classical ML, Deep Learning (CNN, RNN), Natural Language Processing (NER, Sentiment Analysis), Reinforcement Learning, Computer Vision (Object Detection, Segmentation), Predictive Analytics, Time-Series Forecasting

LLMs & Generative AI:

Fine-tuning (LoRA/PEFT), RAG, Agentic AI, Diffusion Models, GANs, VAEs, Prompt Engineering, Evaluation

Libraries & Frameworks:

Pandas, NumPy, PySpark, Scikit-learn, XGBoost, PyTorch, TensorFlow, HuggingFace Transformers, Tokenizers, Ollama, LangChain, LangGraph, AutoGen, Semantic Kernel, CrewAI, OpenAI, Azure AI, OpenCV, NLTK, Spacy, Gym, Streamlit, Flask, FastAPI, Matplotlib, Plotly, PowerBI

MLOps:

Git, ArgoCD, MLflow, TensorBoard, GitHub Actions (CI/CD), Azure DevOps, Docker (Containerization), Kubernetes

Cloud Platforms:

Microsoft Azure, AWS (Basics)

Data Engineering:

ETL/ELT, Data Lake / Lakehouse, Databricks, MySQL, PostgreSQL, Vector Databases (Azure AI Search, FAISS, ChromaDB, Pinecone)

Data Engineering:

ETL, Databricks, MySQL, PostgreSQL, Vector Databases (Azure AI Search, FAISS, ChromaDB, Pinecone)

Soft Skills:

Stakeholder Communication, Product Thinking, Analytical Reasoning, Cross-functional Collaboration

PROJECTS

LLM-Driven AI Copilot

• Developed a multi-modal AI co-pilot using a fine-tuned LLaMA model with PEFT (LoRA), Whisper ASR, and vision-language-models for real-time ATC communication and in-flight decision-making.
• Deployed a RAG pipeline (LangChain, FAISS, HuggingFace) to retrieve and contextualize emergency protocols, enabling accurate and timely autonomous responses.
• Integrated RL-based control agents with LLM agents in LangGraph to enable context-aware autonomy, supporting human-in-the-loop overrides and enhancing overall flight safety.

RAG Pipeline for Semi-Structured PDF Querying

• Developed a RAG pipeline using LangChain, ChromaDB, and HuggingFace Transformers to extract, summarize, and query text and tabular data from semi-structured PDFs via a Streamlit-based UI.
• Implemented chunking, summarization, vector embedding, and multi-vector retrieval, enabling accurate semantic search and real-time LLM responses, with local caching to eliminate redundant computation.
• Integrated OCR (Tesseract) and table parsing (Poppler, Unstructured library) for handling scanned PDFs and complex tables, significantly improving data accessibility and reducing query latency.

Scalable Movie Recommendation Platform (MERN Stack)

• Designed and developed a production-ready movie recommendation website using the MERN stack with secure authentication and RESTful APIs.
• Implemented personalized movie recommendations using collaborative filtering and content-based algorithms, improving user retention and engagement.
• Integrated Google Ads and analytics for monetization and performance tracking, supporting scalable deployment.