AI/ML Engineer
Summary
Aspiring AI/ML Engineer focused on building real-world GenAI and machine learning applications.
Technical Skills
- Languages: Python, SQL
- GenAI & LLM: LangChain (RAG), Langsmith, Groq API, Google Gemini API
- ML/DL: TensorFlow, Keras, Scikit-learn, neural networks, Transformer models, transfer learning
- MLOps & Deployment: MLflow, DVC, Docker, Kubernetes, CI/CD (GitHub Actions), AWS (EC2, S3, ECR)
- Databases: MongoDB Atlas, MySQL, VectorDB (Pinecone, FAISS)
- Other Tools: FastAPI, Streamlit, Git, PowerBI
Education
VIT Vellore
M.Sc. Data Science - CGPA - 8.38
July 2024 – Present
Vellore, Tamil Nadu, India
Atmiya University
Bachelor of Computer Application - CGPA - 8.79
June 2021– April 2024
Rajkot, Gujarat, India
Data Science Training Program
GeeksforGeeks
Jan 2024 - March 2024
Remote
- Completed an intensive 12-week certification course with GeeksforGeeks, mastering Python and its key libraries including Pandas, NumPy, Matplotlib, and Seaborn.
- Familiar with EDA, feature engineering, supervised/unsupervised learning model Classification/Regression algorithm, ensemble techniques (i.e., XGBoost, Random Forest), hyperparameter tuning, and clustering algorithms (K-means, DBSCAN).
- More emphasis on Excel, SQL, Power BI, web scraping using Beautiful Soup and Selenium, and deployment via Streamlit.
Work Experience
Ignite Intern
Machine Learning Intern
May 2024 – June 2024
Remote
- Built a customer segmentation model using RFM (Recency, Frequency, Monetary) features and K-Means clustering on the Online Retail dataset.
- Cleaned and transformed raw transaction data into RFM scores to understand how recently, how often, and how much each customer buys.
- Analyzed the clusters to identify patterns in customer purchase behavior(loyal, high-value, and low-engagement customers).
Jupical Technology
Python Technical Training
April 2023 – May 2023
Rajkot, Gujarat
- Got hands-on experience in Python development concepts, debugging techniques, and logical problem-solving.
- Built a basic e-commerce website using HTML, Bootstrap, and JavaScript, and implemented the backend with Django and SQLite.
Research Work
Garbage Segregation | Python, Transer Learning, Tensorflow, Keras, Matplotlib, Seaborn, Pandas, Numpy
Github Link
- Designed and implemented a multimodal deep learning model that combines image features using CNN and synthetic gas sensor data using MLP to classify waste into biodegradable and non-biodegradable categories (merged from 12 original classes).
- Analyzed each waste class, preprocessed and cleaned the data and applied data augmentation and feature engineering to enhance model performance, achieving 97% training accuracy and 95% validation accuracy, and validated the model’s performance on real-world images.
- Authored and presented a research paper detailing methodology, experiments, and results at a VIT-AP International Conference.
Projects
AI-Powered Document & Data Chat Platform | Python, LangChain, Google GenAI, FastAPI, Pinecone, MongoDB Atlas, Docker, Streamlit
Github Link
- Built a Generative AI chat system for natural language queries on CSV, Excel, PDF, and DOCX files.
- Integrated LangChain with Google Gemini API for retrieval-augmented generation (RAG).
- Used FastAPI for backend with document upload/retrieval/summarize APIs and MongoDB Atlas for persistent user-specific chat history.
- Used Pinecone vector database for semantic search, Implemented async/await concurrency in FastAPI backend to handle multiple simultaneous user queries without blocking I/O.
- Containerized backend & frontend into a single Docker image for seamless deployment.
MLOps- Next Word Predictor |LSTM, MLflow + DagsHub, DVC, Docker, AWS, CI/CD, Flask
Github Link
- Built an MLOps pipeline integrating AWS(S3) and CI/CD with GitHub Actions, automating the entire deep learning workflow from data ingestion to model registration.
- Scraped text data from the web using BeautifulSoup and Requests for data creation.
- Trained an LSTM model on text data and deployed it using Flask containerized with Docker.
- Implemented robust logging and exception handling using Python’s logging module for smooth debugging and error tracking.
MLOps- Sentiment Analysis | Machine Learning, MLflow + DagsHub, DVC, Docker, CI/CD, Flask
Github Link
- Built an MLOps pipeline integrating DVC, MLflow, GitHub Actions, and DagsHub automating the entire sentiment analysis workflow from data ingestion to model registration.
- Implemented CI/CD workflow with GitHub Actions to automatically run pipeline stages, model validation, and API tests on code commit.
- Managed experiment tracking through MLflow to monitor model metrics, compare performance across versions, and streamline model selection.
- Trained a RandomForestClassifier on text data and deployed it via Flask API containerized with Docker and implemented robust logging and exception handling using Python’s logging module for debugging and tracking.
Certifications
- Presentation certificate- ”Garbage Segregation Using Deep Learning”, VIT-AP International Conference
- GeeksforGeeks- Data Science Training Program Certificate
- Certificate of Code Carnival- An Open Hackathon [36 h]
- Certification of completion of basic technical Python training
- Certificate of Academic Excellence in Advanced Java