Vipul Pai Raikar

Algorithms and system integration team lead for next generation VeriSight Pro intra-cardiac imaging catheter. https://www.usa.philips.com/healthcare/product/300000239661/verisightpro3dicecatheter

Product Owner & Team Lead on AI driven system automation projects. - Research Lead & Product Owner for the General Imaging Ultrasound Auto‑Label and Auto‑Measure program, leading a globally distributed team (Bangalore, Shanghai, Cambridge) and coordinating closely with Ultrasound Business stakeholders (Cambridge, Bothell) and external partners (Bangalore) for data curation and annotation. - End‑to‑end ownership of AI system delivery, including data annotation pipeline setup and quality review, prototype integration and debugging, in‑house validation scans, milestone reporting, and coaching junior scientists and interns. - Delivered on‑cart, real‑time AI deployment on EPIQ ultrasound scanners under aggressive timelines for RSNA 2023—collaborating across R&D and business teams to integrate Auto‑Label and Auto‑Measure models, achieve live scanning milestones, and meet safety tested EPIQ software release shipment deadlines. This was used to perform live scan at RSNA on models. - Drove significant business impact, passing 4 major milestones (CS and CP) in 2023, enabling successful RSNA NDA‑room demonstrations with strong customer and KOL validation, and influencing leadership decisions to accelerate product timelines by ~1 year. - Recognized for technical excellence and leadership, including public acknowledgment by GI Business and Clinical leaders for RSNA 2023 success, strong customer resonance during live demos, and multiple leadership shout‑outs for ownership, execution under pressure, and cross‑team collaboration.

Philips Research North America | IGT Integrated Procedures Group Technical Lead & Product Owner - Designed and implemented a hardware‑optimized AI inference module for first-of-its-kind real-time 6DoF implant pose estimation in 3D ultrasound, targeting NVIDIA GPUs with system‑level APIs (CUDA, TensorRT) and achieving a 5× inference speedup (3→15 fps) through architectural and GPU optimizations. - Optimized AI architectures using parallelized and dilated convolutions, enabling network compression and faster execution; delivered a modular C++ library with stable APIs, wrapped in C# for seamless integration into Philips EPIQ platform. - Built scalable annotation and data curation pipelines for large, procedure‑specific clinical datasets; developed multiple generations of annotation tools that reduced labeling time by >50% and established clinically validated data processing strategies. - Evolved annotation methodologies from active‑contour segmentation to model‑based fitting, improving AI training performance by focusing on clinically relevant features; delivered an extensible, user‑friendly annotation platform (Python, Slicer3D). - Led a cross‑functional team to design and develop an X‑ray image segmentation algorithm in 5 weeks, capable of accurately segmenting central lines in highly cluttered X‑ray images, achieving a top‑3 finish in a global, company‑wide challenge spanning four continents. - Help build a strong IP portfolio with 25+ patent applications.

Bio-instrumentation and Electronics

Designed and implemented a tracked freehand point-of-care ultrasound system and associated algorithms for kidney volume measurement, enabling accurate assessment of large kidney volumes within 10% of reference standards—significantly improving upon the ~25% error observed with conventional methods. Expertise spans 3D tool tracking, image registration, statistical shape and appearance modeling, and computer vision applications in ultrasound imaging.