AI & ML Engineer — building end-to-end pipelines, predictive models, and LLM applications that drive measurable real-world impact.
📍 New York, NY · ✉️ amey.borkar01@gmail.com · 🔗 LinkedIn · 🔗 Medium · 🐙 GitHub
- Engineer predictive demand and reorder-forecasting models using XGBoost, Random Forests, and time series techniques (ARIMA, Prophet) on multi-source logistics datasets, improving supply chain accuracy by 25%
- Prototype Graph Neural Networks (PyTorch Geometric, NetworkX) to uncover hidden relational patterns and improve lead time estimates across logistics networks
- Automate data ingestion and transformation pipelines with Python & Pandas, reducing manual prep time by 30%
- Develop an LLM-powered enterprise chatbot enabling natural language querying over 150K+ internal logistics records
- Managed instruction, mentoring, and grading for 180+ students across 3 graduate courses (Mathematical Foundations of Analytics, Python Programming, Artificial Intelligence)
- Created 20+ hands-on exercises and real-world case studies, contributing to a ~15% improvement in average exam scores
- Designed, trained, and validated ML & deep learning models on 50,000+ GC/MS reports, improving formulation accuracy by 20%
- Deployed a scalable end-to-end ML pipeline on Azure ML Studio with MLflow for experiment tracking and model versioning
- Conducted A/B testing and partnered with cross-functional R&D teams to integrate insights into flavor development workflows
- Developed XGBoost-based loan default prediction models on 800K+ records with domain-informed feature engineering, improving high-risk detection by 20%+
- Deployed model outputs into a live Power BI dashboard with integrated SHAP explanations, enabling real-time underwriter risk insights
| Project | Description | Tech |
|---|---|---|
| Flamingo Cares | AI medical training simulator with virtual patient chatbot | AWS Bedrock, Claude Sonnet 3.5, Textract, Titan Embeddings, S3 |
| Teach Simple | Real-time voice-driven educational assistant — 40% increase in user engagement | OpenAI Realtime API, Whisper, Flask, WebSockets |
| GreatVillage | AI property management platform | Next.js, Tailwind, LangChain, LiveKit, Groq |
| Memory Palace CLI | Terminal study assistant — flashcards, MCQs, mnemonics & smart analytics | Python, LLMs |
| Attrition Forecasting | ML model & Streamlit UI for employee job-switch prediction | Scikit-learn, Streamlit |
| Flight Delay Analysis | Big data pipeline & ensemble models — reduced manual review by 30% | PySpark, Hive |
| Virtual Fencing | Real-time CV-based trespassing detection | OpenCV, YOLOv8 |
| Laptop Price Prediction | EDA + Random Forest → 93% accuracy | Python, Scikit-learn |
- 🥇 1st Place — Voxel51 Visual AI Hackathon: CV-based virtual fencing for railway safety (YOLOv8, FiftyOne)
- 🥉 3rd Place — Hatch Labs Hackathon: Intelligent community management platform (LangChain, Agentic RAG)
- 🏅 Best Use of Linkup — Datadog Hackathon: Smart relocation assistant (Linkup Web Crawler API)
MS in Data Science — Pace University, New York, NY (May 2025) Scholar Award | GPA: 4.0/4.0 | Lead, Pace Data Science Club
BE in Electronics & Telecommunication — YCCE, Nagpur, India (Jun 2022) Honor in Automation and Computer Vision | GPA: 9.17/10.0
- 📄 Prediction & Classification of Diabetes Diseases — IEEE BECITHCON 2021 / IEEE Xplore (Random Forest: 98%, SVM: 92% on 3,500+ hospital records)
- 📄 Fake News Classification — Springer (NLP-based detection, 96% accuracy on 8,000+ articles)
- 🎓 Microsoft: Azure Fundamentals (AZ-900), Azure Data Scientist Associate (DP-100)
- 🏅 AWS: Certified Solutions Architect – Associate (2025)
- 💻 IBM: Python 101, Data Analysis with Python, Machine Learning with Python, Data Visualization with Python
- 🛠️ Forage: PwC Power BI, British Airways Data Science
✉️ amey.borkar01@gmail.com · 🔗 LinkedIn · 🔗 Medium · 🐙 GitHub