Characterising C&D waste using Computer Vision
→
Summary
Developed an automated computer vision (CV) system capable of processing 2,000+ images daily, significantly boosting waste detection precision by 20%.
About
Highly accomplished AI-Driven Full Stack Developer with 1+ years of experience, specializing in fine-tuning LLMs and Transformer models for production-ready inference pipelines. Expertly bridges AI and engineering, leveraging MLOps, GCP, and full-stack proficiency in Next.js, React, and FastAPI to deliver scalable, real-time, user-centric AI and web solutions. Proven ability to architect and deploy high-performance systems, driving significant improvements in user satisfaction, operational efficiency, and data-driven decision-making.
Work
Bosswallah Technologies Pvt. Ltd.
|Fullstack - AI Engineer
Bengaluru, Karnataka, India
→
Summary
Led full-stack AI engineering initiatives to develop and scale user-centric platforms, significantly enhancing performance and user engagement for Bosswallah Technologies.
Highlights
Architected and deployed an AI-powered feedback analysis system, boosting customer satisfaction scores by 30% and enabling data-driven strategic decisions.
Spearheaded the launch of the BossWallah platform using Next.js, FastAPI, and TypeScript, achieving scalability for 10,000+ monthly users and reducing onboarding time by 40%.
Engineered and scaled the Expert Connect system with FastAPI, consistently supporting over 10,000 monthly active users with high-performance and reliability.
Integrated robust Razorpay and Firebase authentication, ensuring 100% secure transactions and decreasing failed login attempts by 25%.
Optimized RESTful APIs and refined client-server communication protocols, resulting in a 35% improvement in application load times and overall system performance.
Suvision Holdings Pvt. Ltd.
|AI ML Engineer
Bengaluru, Karnataka, India
→
Summary
Engineered and deployed scalable AI/ML pipelines for sales call analysis, driving significant automation and data efficiency for Suvision Holdings.
Highlights
Deployed a Sales Call Analysis Pipeline, processing over 50,000 records monthly and generating strategic insights using LLMs, GenAI, ML, and DL models.
Architected and implemented scalable AI workflows using Python, RAG, Generative AI, and MongoDB/ClickHouse, achieving over 90% data accuracy.
Automated the entire ML pipeline from data ingestion to storage, eliminating 75% of manual tasks and accelerating data processing time by 50%.
Integrated GCP Storage and BigQuery to enhance batch processing efficiency by 50%, ensuring 24/7 data availability and reliability.
Implemented full automation for insights extraction using AI/ML models, significantly reducing manual effort and improving responsiveness by 75%.
Bitwise
|Data Science Internship
Chennai, Tamil Nadu, India
→
Summary
Developed and deployed AI/ML systems and applications, enhancing content generation and predictive analytics for Bitwise.
Highlights
Devised a GAN-based content generation system for enterprise applications, successfully increasing creative asset output by 30%.
Built an end-to-end Student Performance Prediction System leveraging Random Forest and Flask, achieving over 92% prediction accuracy.
Developed and deployed a Flask application for real-time ML predictions, improving system usability and reducing average response time by 40%.
Containerized the Flask application with Docker, ensuring 100% environment consistency and mitigating deployment issues by 80%.
Education
Indian Institute of Technology, Madras
Chennai, Tamil Nadu, India
→
Bachelor of Technology (B.Tech)
Technology
Skills
Languages
JavaScript, TypeScript, Python, Java, C++, C, SQL.
Web & Frameworks
React.js, Next.js (SSR), Node.js, Express.js, FastAPI, Flask, TailwindCSS, Material UI, Bootstrap.
Databases
MongoDB, MySQL, ClickHouse, BigQuery, Database Management Systems (DBMS), Relational Database Management Systems (RDMS).
AI/ML
TensorFlow, Keras, PyTorch, HuggingFace, Transformers, LLMs, RAG, CNNs, Scikit-learn, OpenCV, Apache Spark, NumPy, Pandas.
DevOps & MLOps
Docker, CI/CD, Git, GCP, JIRA, MageAI, MLOps, Inference Pipelines.
Other
RESTful APIs, Microservices, Redux, Context API, System Design, Data Structures and Algorithms, Agile, Server.
Projects
Human activity Recognition
→
Summary
Designed a CNN-based activity recognition pipeline, processing 5,000+ frames daily with TensorFlow and Keras, achieving over 92% accuracy.