Machine Learning Engineer • MLOps • NLP • LLM Systems
Currently working as a Machine Learning Engineer at Ada IQ, Inc. — building production-grade ML/NLP pipelines, from entity resolution over 7M+ records to RL-guided knowledge graphs and multi-modal evaluation models deployed on SageMaker.
- Architecting end-to-end ML systems — data pipelines, model training, serving, and monitoring
- Experienced with LLM fine-tuning, agentic AI, knowledge graphs, and multi-modal deep learning
- Strong focus on productionizing ML: Docker, Kubernetes, Terraform, CI/CD, AWS infrastructure
Ada IQ, Inc. — Machine Learning Engineer Co-op | Boston, MA | Jan 2025 – Ongoing
Built and productionized the entire ML pipeline end-to-end for a Fortune 500 apparel client
- Productionized entity resolution (Glue/PySpark; adaptive salting + trigram/Jaccard) over 7M+ entities; reduced duplicates by 20%
- Architected RL-guided knowledge graph (Neo4j) for product concept generation using PPO with graph traversal agent
- Built multi-modal evaluation model (ResNet-50 + DistilBERT) with self-attention fusion; deployed on SageMaker
- Automated infrastructure with Terraform + GitHub Actions orchestrating AWS Glue, Lambda, and Step Functions
- Productionized BERTopic with Bayesian HPO; shipped SQL-ready topic features in Athena joined with ABSA + sentiment
- Built dual-mode ABSA pipeline (guided + unguided) using Gemini with Function Calling on 4M+ comments
- Identified critical performance drivers from 1M+ multi-channel comments using ABSA and statistical tests
Quantiphi Inc. — QA Automation Engineer | Mumbai, India | Oct 2022 – Aug 2023
- Automated data extraction/validation in Python/SQL for Workday rollout; built KPI dashboards improving efficiency by 25%
Infosys Ltd. — System Engineer (SDET) | Pune, India | Nov 2020 – Oct 2022
- Developed end-to-end test automation with Java, Selenium, Cucumber (BDD); reduced test cycle time by 40%
Languages
ML / Deep Learning / NLP
LLMs & AI Agents
MLOps & Cloud
Data & Backend
|
Production ML pipeline for Llama 3.2 fine-tuning with LoRA/QLoRA, FastAPI + vLLM serving, Docker, Kubernetes, MLflow tracking, and CI/CD. End-to-end from data preprocessing to deployed inference.
|
Multi-agent AI system using Gemini Flash API that autonomously diagnoses and fixes code bugs through an iterative observe-plan-act-verify loop with Planner and Reviewer agents.
|
|
MCP-based AI agent for real-time ML model monitoring & drift detection (KS-Test, PSI, Wasserstein) with LangFlow visual pipeline, LLM root cause analysis, and K8s integration.
|
Company similarity framework using FAISS, Sentence Transformers, TF-IDF, BM25 with Gemini-powered summaries; deployed via FastAPI + Streamlit.
|
|
Automated GTM validation pipeline using LLM framework to generate personas and rank investors. Won 2nd place at MIT AGI House × AI21 Labs hackathon.
|
Full-stack Carbon Credit Exchange built from scratch — MySQL/MongoDB database design, 11+ advanced SQL query patterns, interactive Streamlit dashboard with Plotly analytics.
|
🎓 Northeastern University, Boston, MA — MS, Data Analytics Engineering
Machine Learning, NLP, MLOps, Deep Learning, Data Mining
🎓 NMIMS (Mukesh Patel School), Mumbai, India — BTech, Computer Engineering
Data Structures, Algorithms, AI, Databases
Open to opportunities in ML Engineering, Data Scientist, MLOps, NLP, and LLM Infrastructure roles.
📍 Boston, MA | 📧 ghatage.r@northeastern.edu