2018
Integrated MSc Data Science
Coimbatore Institute of Technology
2018 - 2023
Open to full-time data and AI roles
Theepan Kumar Gandhi · AI / ML EngineerAI / ML Engineer
Building AI systems that are thoughtful, reliable, and ready for production.
Focus
Generative AI
→
Agentic Workflows
→
Retrieval Systems
→
MLOps
→
→
Production AI
AI Engineer building machine learning products, retrieval systems, and data applications that turn complex information into usable decisions.
0+
Projects Built
0%
Top-k Retrieval
0+
Years Exp
0
Certifications
About Me
Most ML projects fail after the model works. The hard part isn’t accuracy — it’s making the system reliable enough to actually ship. That’s the problem I care about: building AI that holds up in production, not just in notebooks.
I’m a Machine Learning Engineer and recent Master’s graduate from Illinois Institute of Technology, Chicago. Over the past few years I’ve built across the full ML stack — from RAG pipelines and multi-agent systems to recommender engines and multimodal search. My practicum at Pure Platform is a good example: a hybrid text-image search engine that hit 92% top-k retrieval accuracy, not by tuning one component but by getting the embedding strategy, FAISS indexing, and reranking pipeline to work together under real query load.
What I enjoy most is the architecture layer — deciding how models, memory, retrieval, and orchestration connect so the system actually behaves predictably. I work primarily in Python, PyTorch, Hugging Face, LangGraph, LlamaIndex, and FAISS on the ML side, and AWS, Docker, Kubernetes, and Terraform when it comes to deployment.
I’m currently open to full-time roles in ML Engineering, AI Engineering, or Applied Science — anywhere the work involves taking intelligent systems seriously from prototype to production.
25+Projects Built
3Years Experience
6Certifications
92%Top-k Retrieval
Skills & Expertise
Move your cursor inside — the graph reacts to you
Education & Experience
2021
Machine Learning Intern
Optisol Business Solutions, India
Jun 2021 - Nov 2021
2023
Machine Learning Intern
Peculiaar Automation Solutions, India
Jan 2023 - May 2023
2024
Master's in Data Science
Illinois Institute of Technology
2024 - 2025
2025
Data Science Intern
Pure Platform, Chicago, USA
Jan 2025 - May 2025
2018
Coimbatore Institute of Technology
Integrated MSc Data Science
- Built a strong foundation in Data Structures, Algorithms, Probability and Applications, Business Statistics, and Multivariate Data Analysis, strengthening analytical and problem-solving skills for data science.
- Developed solid systems and technical depth through Computer Networks, Operating Systems, Machine Learning, Deep Learning, and programming in Python and R.
- Strengthened applied data science expertise through Big Data and Data Mining coursework, supported by practical projects focused on scalable, data-driven solutions.
2021
Optisol Business Solutions, India
Machine Learning Intern
- Engineered a real-time data prefetcher to acquire and stage time-series network telemetry streams into Amazon S3 for ML prediction.
- Developed and deployed a Prophet time-series forecasting model on SageMaker, exposing a real-time prediction endpoint for telemetry data.
- Deployed FastAPI endpoints to expose Prophet SageMaker predictions, enabling real-time network telemetry forecasting for enterprise clients.
- Optimized MongoDB Atlas data workflows, improving retrieval speed by 30% and enabling scalable analysis in a high-performance database.
2023
Peculiaar Automation Solutions, India
Machine Learning Intern
- Developed a deep learning attendance system using Convolutional and Siamese Networks for accurate facial recognition and identity matching.
- Automated nearly 70% of manual attendance tracking, significantly improving operational efficiency for HR workflows.
- Integrated MySQL for secure record management, reliable attendance storage, and streamlined access control.
- Improved system usability by enabling faster attendance capture and reducing manual errors in day-to-day workforce tracking.
2024
Illinois Institute of Technology
Master's in Data Science
- Completed advanced graduate coursework in Machine Learning, Artificial Intelligence, and Deep Learning, building strong expertise in modern AI and predictive modeling.
- Strengthened analytical skills through Regression, Statistical Learning, and Monte Carlo methods for rigorous model evaluation and data-driven decision making.
- Built solid data and AI foundations through Big Data, Database Management, and Data Preparation & Analysis, with practical application in scalable machine learning solutions.
2025
Pure Platform, Chicago, USA
Data Science Intern
- Engineered and served a multimodal hybrid product search engine supporting simultaneous text and image input, leveraging FAISS for low-latency vector similarity search and achieving 92% top-k retrieval accuracy, helping users find the right product faster.
- Fused OpenAI CLIP-based text and image embeddings into unified vectors, incorporating BM25 scoring to improve text understanding and image relevance, enhancing multimodal search quality and user satisfaction.
- Optimized search with FAISS IndexHNSW, achieving 20% faster retrieval and 2x concurrent searches without additional costs.
- Structured a flexible search pipeline supporting image uploads, product URLs, and keyword queries, enabling API integration for real-time retrieval and boosting relevance by 25% via weighted alpha blending, improving product discovery for users.
What I Do
Machine Learning Solutions
Designing and deploying custom ML models for predictive analytics, classification, and intelligent automation.
Data Analytics & Visualization
Transforming raw data into actionable insights using dashboards and visual storytelling with Power BI and Tableau.
NLP Applications
Building NLP pipelines for tasks like text classification, semantic search, chatbots, and information extraction.
Recommendation Systems
Developing personalized recommendation engines using collaborative filtering and deep learning approaches.
Model Deployment & MLOps
End-to-end deployment of ML models using Docker, Kubernetes, and AWS with scalable pipelines.
Vector Search & Semantic Retrieval
Implementing vector-based search using FAISS or Pinecone for similarity search and Retrieval-Augmented Generation (RAG).
Data Preprocessing & Feature Engineering
Cleaning, transforming, and engineering features to maximize model performance and business relevance.
API & Web App Integration
Integrating ML models into web apps using Flask, FastAPI, or Streamlit for real-time inference.
Recent Projects 20
Terraform-Based LLMOps Platform (Canary Deployment)
Evaluation-gated canary deployment of LLMs using Terraform infrastructure-as-code.
View on GitHub →
Autonomous Agent for GitHub Actions CI Fixer
LLM agent that autonomously detects and fixes failing GitHub Actions CI pipelines.
View on GitHub →
Agent Service Platform — Multi Agent Orchestration
Coordinates multi-agent workflows, tools, and service execution across a unified platform.
View on GitHub →
Finance Document Assistant (RAG + Agents)
RAG + agent pipeline for accurate financial document question answering.
View on GitHub →
Orchestration: End-to-End AutoML Pipelines
Graph-based agent orchestration for automated end-to-end ML workflows.
View on GitHub →
Malicious URL Detection
ML pipeline for detecting malicious URLs and suspicious network patterns.
View on GitHub →
Data Science Notebook Assistant (QLoRA Fine-tuning)
QLoRA fine-tuned LLM providing theory and code guidance in data science notebooks.
View on GitHub →
Data Quality Profiler
Automated profiling tool to detect schema drift, missing values, and statistical anomalies.
View on GitHub →
LLM Annotation Quality Pipeline
Pipeline for evaluating and improving annotation quality in LLM fine-tuning datasets.
View on GitHub →
Two-Stage Recommender (FAISS + LightGBM)
FAISS candidate retrieval (Top-200) followed by LightGBM reranking (Top-10).
View on GitHub →
Multi-Stage Two-Tower Recommender
Two-tower retrieval model with downstream reranking for personalized recommendations.
View on GitHub →
Gmail Automation with CrewAI
Agent-driven automation for handling and organizing routine Gmail operations.
View on GitHub →
Multimodal RAG Assistant (LlamaIndex)
Multimodal retrieval assistant for grounded Q&A across text and image inputs.
View on GitHub →
RAG Powered QA Chatbot
Question-answering chatbot grounded in retrieved document context to reduce hallucinations.
View on GitHub →
Credit Risk Prediction with LightGBM, CatBoost & SHAP
Interpretable credit risk model with gradient boosting and SHAP explainability.
View on GitHub →
ECG Anomaly Detection
LSTM autoencoder for detecting abnormal heartbeat patterns in ECG signals.
View on GitHub →
Music Genre Classification
Deep learning model for audio signal-based music genre classification.
View on GitHub →
Stock Forecasting with LSTM + ARIMA
Hybrid LSTM + ARIMA pipeline for robust stock price time series forecasting.
View on GitHub →
Diabetes Disease Progression Prediction
Regression model with SHAP explainability for diabetes progression trend analysis.
View on GitHub →
Multimodal RAG Assistant with Chat History
RAG assistant with multi-input support and persisted conversational memory.
View on GitHub →