Prasanna Thirukudanthai Raghavan

About Prasanna Thirukudanthai Raghavan

Hello! I\'m Prasanna Thirukudanthai Raghavan, a motivated Robotics Engineer with a passion for leveraging technology to create innovative solutions. With expertise in Robot Perception, Computer Vision, and Controls, I thrive on pushing the boundaries of robotics and automation.I have a proven track record in cutting-edge projects, working on autonomous systems, computer vision algorithms, and intelligent control systems. Proficient in Python, C++, and MATLAB, along with frameworks like ROS, OpenCV, TensorFlow, and PyTorch, I develop state-of-the-art solutions that optimize performance and drive innovation.Beyond technical skills, I excel in collaboration, working effectively within multidisciplinary teams and communicating complex concepts. I\'m fueled by ambitious projects and a desire to learn from and inspire others.I am an enthusiastic Robotics Engineer, seeking to make a meaningful impact through innovation. Let\'s connect to explore how we can drive transformative change together in robotics and automation.

Experience

1. Ai Software Engineer

   Rts Labs

   Jun 2024 — Present · US

   Intelligent Edge Data Pipeline: Built an LLM pipeline using LangChain, GPT-4, and Python to extract REDCap data from multi-type forms. Improved data consistency by 95% and query speed by 40%, saving 100 hours/month ($18K/quarter) via automated schema mapping and JSON restructuring. CAD-Based Part Identification: Developed a BOM automation system using Python, OpenCV, and Bluebeam’s Revu APIs to extract parts from 250+ annotated CAD drawings.Simulated real-time object identification and structured data generation, mimicking automated robotic recognition in industrial workflows. Delivered a 30% faster turnaround on bids, improving estimation consistency. Auction Item Scheduling: Built an AI-driven scheduling engine using OpenAI GPT-4, Pinecone embeddings, and FastAPI to recommend daily auction lots ( 9.5K+ items). Reduced scheduling latency by 70% and boosted lot engagement using pattern learning, time-series inference, and external signal analysis. Object Detection Viability Study: Benchmarked YOLOv8, Detectron2, Grounding DINO, OWL-ViT, and CLIP using Python, PyTorch, and Roboflow. Found 92% mAP in traditional models vs. 85–88% in VLMs; evaluated feasibility for edge deployment (Jetson Nano, Tegra). Analyzed GPT-4V, LLaVA, and BLIP-2 for zero-shot detection and visual grounding to inform 2025 AI product strategy.

Education

• University of Maryland

  Master's In engineering, Mechatronics, Robotics, and Automation Engineering

• SRM IST Chennai

  Bachelor of Technology - BTech, Mechatronics, Robotics, and Automation Engineering