Ashok Dahal, PhD

• Advancing multi-class perception systems to detect diverse 3D objects (vehicles, pedestrians) and complex lane topologies, fusing Camera, LiDAR, and Radar data to power L2/L3 ADAS features. • Benchmarking exterior perception performance against Ground Truth systems, maintaining rigorous validation protocols to guarantee ongoing robustness for safety-critical applications. • Engineering scalable data pipelines that ingest and process raw LiDAR and image streams, aggregating structured perception metrics in BigQuery to drive dashboard-based requirements validation. • Iteratively refining ML models through advanced data analysis and visualization loops, continuously fine-tuning hyper-parameters to maximize detection accuracy and system reliability.

• Computer Vision Feature Owner for various ADAS features such as Trailering Assistance, Automated Parking and Camera Lens Blockage Detection. • Valeo Expert on Computer Vision and Machine Learning. • R&D of vision based Perception Algorithms for ADAS features supplied to North American Automotive manufacturers. • Involved in Intellectual Property(IP) clearance and patent filing for various Computer Vision and Machine Learning Algorithms.

• Developed and optimized Lane Sensing algorithms in C++ for Surround View Cameras integral to GM's Super Cruise feature. • Conducted research and development on camera-based Perception Algorithms including Object Detection and Tracking, contributing to the advancement of Advanced Driver Assistance Systems (ADAS). • Played a key role in routine software design reviews, peer code reviews, and software quality test activities, ensuring high standards of reliability and performance. • Analyzed and refined algorithms for improved accuracy and efficiency, enhancing the overall functionality and safety of ADAS features.

• Lab Instructor for Programming Fundamentals I & II using C and C++.

• R&D of Computer Vision and Machine Learning Algorithms for abnormalities detection in colonoscopy videos. • Innovated "DIFF" texture feature extraction algorithm to detect ulcers and abnormalities in colonoscopy videos. • Conducted thorough assessments of texture feature extraction methods (such as LBP, Gabor filters) to refine abnormality detection accuracy. • Proficiently utilized various OpenCV libraries for robust image processing and computer vision solutions.

• Generated electronic wafer maps from raw data collected from testers and probers. • Developed parsers in Perl to extract information from test files. • Managed XML file generation and storage in Oracle database tables. • Designed web reports in Perl/HTML, interfacing with Oracle backend for wafer map generation. • Provided global technical support for electronic wafer map generation.

Dissertation: Detection of Ulcerative Colitis Severity and Enhancement of Informative Frame Filtering using Texture Analysis in Colonoscopy Videos

Specialization - Computer and Network Security

Major - Computer Science Minor - Mathematics