cv

My Resume. You can download it from the button right there 👉

Basics

Name Harshwardhan Fartale
Label Machine Learning & Data Science Professional
Email harshwardhanfartale.nith@gmail.com
Phone +91-9317439486
Url https://emharsha1812.github.io/

Work

  • 2024.07 - Present
    Predoc Associate
    DRDO & Indian Institute of Science, Bangalore
    Working on ML systems for combat aircraft applications and SciML research at AiREX lab under Prof. Sashikumaar Ganesan.
    • Developed ML systems for CASDIC’s combat aircraft applications
    • Evaluated fine-tuned IBM Granite & Gemma models for SciML via Graph-RAG, achieving SOTA results
    • Led Open Source Projects PULSE & SciREX
  • 2024.01 - 2024.06
    Data Scientist
    Bajaj Finserv Health
    Worked on fraud detection models and graph-based pipelines.
    • Built high-performance fraud detection model with best abuse detection rate
    • Developed fraud pipeline using Memgraph, Neo4j, and Cypher
  • 2023.06 - 2023.07
    Machine Learning Intern
    IIT Indore
    Developed ML models for financial forecasting.
    • Predicted stock prices (85% accuracy) using Python & market/news data
    • Enhanced accuracy by 25% via feature engineering & time series analysis
  • 2023.01 - 2023.06
    Open Source Technical Writer
    OpenMined
    Contributed to documentation and technical blogs for privacy-preserving ML.
    • Improved tutorials and user docs for differential privacy
    • Published blogs/articles on federated learning, boosting adoption

Education

  • 2020.12 - 2024.05
    B.Tech
    National Institute of Technology, Hamirpur
    Electrical Engineering

Projects

  • 2025.03 - Present
    AIFred – Secure, Local AI Coding Assistant
    Desktop AI coding assistant with real-time LLM and ASR, ensuring 100% local data privacy.
    • Developed fully local AI pipeline using Ollama & NVIDIA NeMo
    • Voice-command interface for secure offline querying
    • Electron overlay UI for seamless workflow
    • Technologies: Python (FastAPI, NeMo, pydub), Electron, Ollama API
  • Document Tampering Detection
    Forgery detection system using YOLOv8 and Streamlit.
    • YOLOv8 model detecting overwriting, scribbling, whitener use
    • Custom dataset of 600+ document images
    • Streamlit UI for upload & tampering visualization
  • Fundus AI
    AI-powered ophthalmic disease detection solution.
    • Detected Diabetic Retinopathy, Cataract, Glaucoma, ARMD
    • Built deep learning smartphone app for screening
    • 3D-printed Tinkercad accessory for fundus imaging

Skills

Programming Languages
Python
C++
Triton
SQL
ML & Deep Learning
NumPy
pandas
Matplotlib
Scikit-learn
JAX
spaCy
OpenCV
TensorFlow
PyTorch
PySyft
Large Language Models
Huggingface
Ollama
Databases & Query Languages
Neo4j
Memgraph
Cypher
Tools & Platforms
Docker
Kubernetes
LaTeX
MATLAB
Linux Kernel
Google Cloud Platform
Firebase
Adobe Illustrator
Technical Writing
Substack
Medium
Hashnode

Certificates

Volunteer