home_
Machine Learning
Engineer
Focused on computer vision, real-time AI systems,
and data-driven solutions
for social impact
I am Sushil Sharma, a Machine Learning Engineer and Data Analyst specializing in AI/ML and computer vision. I build real-time ML systems with 90%+ accuracy, efficient data pipelines, and FastAPI solutions that cut processing time by 30%.
Languages and Frameworks
Frontend Projects
Technical Projects
ML Composed DDOS Detector
- Implemented a random forest algorithm achieving 89% accuracy in detecting distributed denial-of-service attacks
- Designed seeder and feeder scripts for data pipeline management
- Built three.js designed UI globe for real-time attack visualization
Python
Scikit-learn
Three.js
Machine Learning
Real-Time Drowsiness Detection System
- Combined computer vision techniques with XGBoost algorithm to achieve 94% accuracy detection in real-time conditions
- Published research in ICT-CEEL 2023 Conference and peer-reviewed journal
- Developed alert system for driver safety applications
Python
OpenCV
XGBoost
Computer Vision
Dual Layer Approach for LLM Hallucination Reduction
- Implementing a dual layer approach with pre-generation and post-verification layers to tackle hallucination in Large Language Models
- Exploring novel architectural patterns for improving LLM reliability
Python
PyTorch
LLMs
NLP
Work Experience
ETL Associate
- Built Selenium/BS4 scraping pipelines for automated auction data extraction, achieving 40% efficiency improvement
- Developed Python/Pandas/NumPy ETL workflows for processing and preparing machine learning datasets
- Developed FastAPI APIs with Matplotlib/Seaborn visualization capabilities, reducing data processing time by 30% through modular automation
Python
Selenium
Pandas
FastAPI
BeautifulSoup
Let's Connect
I'm always open to discussing new projects, creative ideas, or opportunities to be part of your vision.
Message sent_