Hello, I'm
Nisyhaal Tamil
I build
|
AI & Robotics Engineer crafting intelligent systems from model development to production deployment
Scroll Down
Hello, I'm
AI & Robotics Engineer crafting intelligent systems from model development to production deployment
Scroll Down
AI and robotics engineer with experience in designing and deploying intelligent systems across the full engineering lifecycle — from model development and optimization to production infrastructure and customer-facing delivery.
Proven track record of building end-to-end solutions spanning edge AI, computer vision, real-time robotics, and generative AI, with deep expertise in bridging research-level techniques to deployment-ready products on resource-constrained hardware.
Driven by a dual ambition: to push the technical frontier in AI and autonomous systems, and to grow into a leadership role that shapes engineering strategy, mentors talent, and steers high-impact product decisions.
Designing, developing, and deploying edge AI solutions spanning computer vision, robotics, generative AI, and model optimization on Intel hardware (Arc GPUs, Xeon, NPUs). Work covers the full engineering lifecycle — from training and optimizing ML models to building production-grade backend services, web frontends, and containerized deployment infrastructure.
Led development of automated AI-based systems for image classification, text classification, and document parsing with focus on business alignment and scalable deployment.
Designed and deployed end-to-end computer vision pipelines for object detection, OCR extraction, and damage identification across edge, on-premises, and cloud environments.
Developed and enhanced computer vision algorithms for segmentation, OCR, fraud detection, and face recognition. Delivered AI-driven solutions and conducted training sessions for teams and stakeholders.
End-to-end agentic AI manufacturing system integrating real-time computer vision, 6-DOF robotic automation, conveyor belt control, and LLM-based natural-language interface for PCB defect detection. Demonstrated at Embedded World 2025.
End-to-end robotic pick-and-place system in ROS 2 Jazzy with YOLO OBB detection achieving 30Hz real-time inference through OpenVINO acceleration. Automated synthetic dataset pipeline using Gazebo Harmonic simulation.
Multi-stage AI inference pipeline chaining 5 distinct models: instance segmentation, YOLOv8 OBB detection, custom OpenVINO OCR with CTC decoding, and RapidOCR fallback with perspective warp rectification.
Multi-modal agentic AI system enabling natural-language querying of visually dense PDF documentation using ColQwen2.5 late-interaction retrieval and Qwen2.5-VL vision-language inference.
Vision-Language-Action framework with ROS 2 industrial gripper driver, real-time MyCobot-to-TM5S teleoperation bridge, and multi-view camera integration for Learning from Demonstration.
Multi-threaded real-time speech translation system running OpenAI Whisper optimized via OpenVINO INT8/FP16 quantization on Intel edge GPUs, supporting simultaneous translation to 3 languages.
Ported 2 modules of a 138K-line C++ optimal control framework (OCS2, ETH Zurich) from ROS 1 to ROS 2 Humble, migrating build systems, C++ APIs, and resolving transitive dependency chains.
Universiti Malaya
Oct 2023 – Jun 2024
CGPA: 3.95
Universiti Sains Malaysia
Sep 2015 – Jan 2020
CGPA: 3.07
Seeking opportunities at the intersection of deep technical ownership and team leadership. Let's build something extraordinary together.