Experience

• 

  Regal systems

  Oct 2019 - Oct 2020

  Game developer

  Role: Unity Game Engine DevelopmentSkills: C# programming, Unity Engine

• 

  Information technology university

  Oct 2020 - Oct 2023

  Role: Data collection, preprocessing and applying machine learning models.Skills: Python, Web Scraping, Numpy, Matplotlib, Pytorch, Tensorflow, scikit-learn Role: Data Acquisition, Research Information for surveySkills: Latex, Python, Writing Research paper Role: Taught labs of Cyber Security Course and labs of Computer Networks.Skills: Python, Socket Programming, Nmap, Ettercap, Wireshark, Nessus, Intrusion Detection, Firewall, DNS, Penetration Testing, Vulnerabilities Analysis, Scanning Networks

  • Research Associate

    Jan 2023 - Oct 2023

  • Research Assistant

    Oct 2020 - Dec 2022

  • Teaching Assistant

    Oct 2020 - Dec 2022
• 

  Xeven solutions

  Mar 2024 - now

  Machine learning engineer

  Projects:1. AI Agent for HealthcareDeveloped an advanced AI agent to enhance healthcare workflows. The agent accepts two key inputs from patients: the chief complaint and the history of present illness (HOPI). The agent performs these core functions: **Symptoms/Medical Terms Extraction**: Automatically extracts all symptoms and medical terms/conditions. **Diagnosis Generation**:Provides a primary diagnosis in compliance with the ICD-10-AM standard, a globally recognized system for classifying diseases and health conditions used in clinical and research settings.**Semantic Search for Medical Information**:Searches medical keywords semantically in a database to deliver relevant descriptions and additional details, aiding healthcare providers in better understanding patient conditions.--->TechStack includes Opensource LLMs(Llama3.1:3b, Llama3.2, Vicuna, Meditron, Llava etc), Commerical LLMs (OpenAI, Med Palm 2 etc), LangChain, LangGraph, LangSmith, Vector Databases (Local-based PgVector, Cloud-based Pinecone,Qdrant), Finetuning Opensource LLMs, FastAPI.2. Realtime Engagement DetectionDeveloped a cutting-edge real-time engagement level detection system to analyze user attentiveness during online sessions or meetings. Core features of project includes: **Eye Gaze Detection**: Implemented gaze direction prediction using a Convolutional Neural Network (CNN) trained on a dataset with three classes (center, left, right), achieving 99.50% accuracy.**Pose Estimation**: Leveraged the pretrained MobileNetV1 model to estimate body poses effectively.**Engagement Analysis**: Combined gaze direction and body pose data to classify user engagement into three levels: fully engaged, partially engaged, and not engaged.---> TechStack includes Deep Learning, Computer Vision(CV), Face Detection, Features Engineering, Python Show less