Durgesh Dutt Sinha
AI/ML Engineer building useful AI products and intelligent interfaces.
I design and ship AI applications, ML systems, and high-performance web experiences.
About Me
I am an AIML Engineer and Full-Stack Architect completing my Master of Computer Applications (AIML) at Sri Balaji University, Pune. As an AI Program Fellow at UNLOXr and Be10x AI Cohort Member, I specialize in building autonomous agents, sub-50ms computer vision pipelines, and resilient full-stack platforms with failover architecture.
Work Experience
AI Cohort Member
Be10x
- ▸ Built autonomous AI agents to optimize real-world workflows
- ▸ Developed full-stack AI product concepts from ideation to demo
- ▸ Utilized data analytics to derive and communicate actionable insights
AI Program Fellow
UNLOX®
- ▸ Hands-on end-to-end AI deployment pipelines
- ▸ Built and tested autonomous AI systems for industry use-cases
- ▸ Industry-level project execution under mentorship
Central Co-Ordination Team
School of Computer Studies – SBUP
- ▸ Volunteer at Freshers Central Coordination Team
Education & Academic Specialization
Formal graduate and undergraduate computer science education combining rigorous Artificial Intelligence theory with production engineering.
Master of Computer Applications (AIML)
Sri Balaji University, Pune•School of Computer Studies
Specialized Master's curriculum focused on engineering scalable AI systems, neural network modeling, and enterprise distributed computing. Applied coursework directly supports research and deployment of autonomous agent pipelines and deep learning web applications.
Bachelor of Computer Applications
Sri Balaji University, Pune
Comprehensive undergraduate foundation in computer science principles, object-oriented design, algorithmic complexity, relational database management, and full-stack software development.
Industry Cloud & ML Credentials
AWS Educate Emerging Talent Community Member
Amazon Web Services (AWS)
AWS Educate Machine Learning Foundations
Amazon Web Services (AWS)
Leadership & Volunteering
Built and maintained web pages for a non-profit organization.
5 Featured Platforms
Curated full-stack AI applications, realtime voice systems, and machine learning pipelines engineered with verified architecture, observable metrics, and production discipline.
RoleRadar — AI Career Intelligence Platform
Full-stack AI resume analyzer and career platform evaluating candidate resumes against 16 industry roles with an 8-point ATS scanner, Google XYZ bullet optimizer, and resilient dual-mode failover storage.
jarvis-realtime-assistant — Realtime Voice AI & Iron Man HUD
Full-stack conversational voice AI system with a futuristic Iron Man HUD. Orchestrates browser Web Audio chunking, Whisper STT, Gemini 2.0 Flash streaming, and edge-synthesized speech with sub-350ms response latency.
MarketMatch-AI — Customer Segmentation & Recommender
End-to-end machine learning pipeline clustering retail consumer behaviors using K-Means and DBSCAN with PCA dimensionality reduction, paired with a Nearest Neighbors recommendation engine for hyper-targeted campaigns.
CNN-STREAMLIT — Deep Learning Computer Vision Classifier
End-to-end Deep Learning Convolutional Neural Network trained on Fashion-MNIST with 89.3% accuracy, featuring an interactive real-time Streamlit image classifier with sketch canvas and photo upload.
FitTrack AI — Computer Vision Fitness Platform
Production-grade 9-page fitness engineering platform featuring real-time MediaPipe pose estimation, kinematic joint angle tracking, exercise rep counting, wearable sensor telemetry, and workout analytics.
Featured GitHub Repositories
Curated engineering selection from public GitHub repositories demonstrating production architectures, distributed systems, and real-world ML pipelines.
CNN-STREAMLIT
End-to-end Deep Learning Convolutional Neural Network trained on Fashion MNIST with 89.3% accuracy and interactive real-time Streamlit image classifier.
Get In Touch
Whether you have an opportunity, an AI architecture question, or want to collaborate on cutting-edge models, I'd love to connect.