Building end-to-end AI solutions for enterprise clients
I lead end-to-end AI project delivery and platform development for enterprise clients. From multi-agent RAG systems to production ML pipelines, I architect intelligent solutions that drive real business value. With 5+ years of experience at industry leaders like Bosch and WAIYS, I specialize in building scalable AI systems from prototype to production.
Design and implement scalable machine learning systems integrated into complete full-stack applications. From data ingestion to model deployment with monitoring and feedback loops.
Transform data into actionable business insights through custom analytics dashboards and business intelligence tools that drive decision-making.
Build robust backend systems and APIs with integrated ML model serving capabilities. Specialized in scalable data infrastructure and real-time processing.
Deploy ML models at scale with monitoring and observability. Implement CI/CD pipelines for continuous model improvement and reliable production systems.
Develop advanced computer vision systems for real-world applications including object detection, tracking, and industrial automation solutions.
Build intelligent assistants and knowledge systems using Large Language Models and retrieval-augmented generation for natural language interaction.
Real-world data-driven applications across diverse industries
Building an enterprise WhatsApp-based AI agent for real-time data querying and business intelligence for a real-estate enterprise client. Overseeing end-to-end development of multi-agent RAG system with 95% accuracy SLA, integrating Monday.com, CRM, and proprietary market intelligence platforms.
Architected a production-grade document processing system using Azure Document Intelligence, OpenAI embeddings, and AI Search to extract and index content from image-embedded PDFs, achieving 95% pipeline reliability with comprehensive cost/latency metrics across 8 integrated Azure services.
Implemented automated semantic versioning for organizational repositories to replace brittle manual tag management. Using a manual-trigger + semantic-release approach where merges to main happen freely while Product Managers control release timing via GitLab CI job.
Built a complete LLM assistant solution leveraging Retrieval-Augmented Generation (RAG) framework with OpenWebUI and asynchronous Python libraries. Designed the full system architecture from database design with ChromaDB to user interface, dynamically generating structured project guides tailored to various team roles.
Architected and prototyped complete integration of various LLMs with proprietary knowledge base systems to provide natural language query results for enterprise teams. Designed the full-stack solution including API layer, vector database integration, and user interface for efficient knowledge discovery and retrieval.
Consulted on implementation of a comprehensive trailer weight detection system using novel neural network architectures and time series analysis. Designed the complete solution pipeline from sensor data ingestion to deployment architecture for accurate weight estimation in production environments.
Implemented a complete multi-modal sensor fusion system fusing data from multiple cameras, generating point clouds and detecting critical area intrusions. Designed the full processing pipeline and integration architecture for enhanced safety monitoring systems.
Configured label detection pipeline for fine-tuning models to classify various light sources during autonomous driving, improving vehicle perception capabilities.
Designed and trained novel neural network architectures for near-field ultrasound sensor-based object classification system, enabling precise autonomous parking capabilities.
Trained object-based detection system to identify visible faults in cell tower components. Deployed on Jetson embedded devices using model quantization with TensorRT for real-time operation.
Implemented state-of-the-art semantic segmentation and object detection models to classify safe drone landing areas. Deployed on Jetson devices using model distillation for optimal performance.
Proven track record delivering full-stack data driven solutions for industry leaders
Leading end-to-end AI project delivery and platform development for enterprise clients. Building multi-agent RAG systems, document intelligence pipelines, and WhatsApp-based AI assistants for real-time business intelligence.
Engineering solutions for cross-industry projects. Designed and implemented complete RAG systems with LLMs, consulted an automobile giant on trailer weight detection system architecture, and designed sensor fusion pipelines for automotive safety systems.
Developed and maintained visual perception pipelines for autonomous systems. Specialized in novel neural network architectures and embedded deployment on NVIDIA Jetson devices.
Specialized in Machine Learning and Deep Learning. Teaching Assistant for ML courses.