Hello World!

• Spearheading enterprise-level Retrieval-Augmented Generation (RAG) system architecture for flagship LLM applications, processing 10M+ lines of code across 15+ programming languages.
• Pioneering cross-functional GenAI initiatives with GCP Cloud, Vector Databases, LangChain, and custom orchestration frameworks, establishing scalable AI solutions company-wide.
• Optimizing development workflows, reducing manual effort by 70% across 25+ client projects, and achieving model performance improvements of 25% in accuracy and 40% in inference speed.

• Automated legacy code refactoring using GCP’s Foundation LLM Models like CodeLlama and Codey, achieving a 95% convergence rate in unit tests.
• Leveraged hybrid prompt techniques for AI model fine-tuning, optimizing retrieval-based tasks and enhancing accuracy for real-time enterprise applications.

• Architected cutting-edge conversational AI system by seamlessly integrating Open Source LLM agents (Llama, Mistral) and Closed Source models (GPT-3.5, GPT-4) with Twilio infrastructure, enabling automated call processing for 500+ daily high-value client interactions.
• Engineered sophisticated Pinecone-based Knowledge Base infrastructure for intelligent call data management and real-time visualization, achieving 90% faster information retrieval and transforming the efficiency of AI calling agents.
• Developed intelligent call routing and response system with advanced context-awareness capabilities, achieving 85% customer satisfaction and significantly improving client retention rates through personalized conversational experiences.

a) Develop and implement technology to detect signed language for effective communication between deaf and dumb individuals.
b) Trained ML models for XR spectacles, enabling seamless communication for sign language users.
c) Utilized Power BI to address data inconsistencies & ensuring balanced representation of hybrid dataset consisting 10,000 ASL and ISL gestures.

a) Designed a ML solution to monitor human vitals, including heart rate, with contactless technology.
b) Using STM32, I integrated this system inside a coffee vending machine, allowing vital tracking with caffeine intake.
c) Resulted in valuable insights about Employee psyche and Caffeine consumption.

a) Hosted AI/ML hackathons
b) An Active speaker in AI/ML speaker series
c) Led the organization and planning of the annual AI/ML research symposium.

a) Organized one of North India’s Largest Hackathon ( Tech-A-Thon 3.0 ) with over 900 participants
b) Mentoring AI/ML and Quantum Machine Learning to students.
c) Organize/Manage various Tech events in Campus.
d) Latest event organized, ( Quantum Machine Learning )

a) Trained and optimized deep learning models for real-time object and facial recognition using diverse datasets like UTK, LFW, and Google’s dataset.
b) Enhanced CCTV surveillance using the trained deep learning models for enhanced tracking.

a) Trained with various ML models to solve real life projects.
b) Mentored various students.
c) Worked upon OpenCv, NLP and sentiment AI like hotcakes.

a) Made a Landing website for the NGO. ( Mangalkari Public Welfare Trust )
b) Monitored ongoing campaigns (Drishti, Sahay)

a) Worked with multi-neural layers.
b) Worked upon the AlexNet model for Image Classification.
c) Integrated AlexNet for medical image classification, as a preliminary test.