• Led team of ML engineers to develop lean DNN model that serves SMT/BGA QC inspection on semiconductor post-processing production line which has extremely low fail rate (<0.01%) and data asymmetry.• Developed DNN model that rates the quality of lithium-ion battery packaging in production line that works in conjunction with existing automated optical inspection system. • Led transition of in-house ML train and serve framework to Ray Train/Tune for simpler code, operational overhead reduction, data parallel training.

• R&D and production design of manufacturing test solutions for Camera, Display, and Illumination (LED/VCSELs) Devices for a major VR/XR vendor with world leading R&D teams in optics, cameras, and sensors.

• Developed DNN solutions for Qualcomm's ML Image Signal Processing (ISP) team from initial concepts to hardware mapping on Snapdragon SoC ASIC. • Developed Qualcomm proprietary camera noise and color models to enhance ML-ISP training to ensure that it can compete with and even exceed the most advanced and sophisticated image processing pipeline designs in the world. • Optimized DNN inference engine for video rendering for integration into Snapdragon SOC. This process entailed custom layer design for memory efficiency, weight/activation quatization, and close collaboration with hardware architectural design team to adopt additional constraints to improve hardware efficiency. Designed phase detection auto focus (PDAF) algorithm from concept to C/C++ firmware and refined commercialized solution until major product releases by tier-1 customers such as Samsung. • Designed phase detection auto focus statistics engine which allows PDAF to also create a dense 3D depth map in real-time.

• Assisted Laboratory Operation in Laser Alignment Setup in Laser Cooling and Trapping Group, Quantum Optics Group, including fabrication of custom parts that maintain vacuum conditions for precise and accurate data collection, leading to improved efficiency and reliable results.• Conducted literature reviews and data analysis, assisting in the development of research proposals and preparing research reports, discussing the latestadvances in the field of Quantum Optics.

• Camera module manufacturing test and calibration protocols with Apple QA engineers.• Full-time studies in optical sciences and full-time supervisory position managing Asia.• Lead rapid manufaturing issue identification and resolutions in East Asia factory operations as QA lead in high-paced, high-capacity production environment.

• Image processing algorithm design including Bayer denoise, multi-exposure HDR, global/local tone mapping, lens shading and lens distortion correction.• HDR algorithms in blending and tone mapping for automotive and security applications.CMOS 4-T, 5-T, 6-T pixel development for mobile, automotive, security, defense applications 3D optical simulation of CMOS pixel devices Image signal processing (ISP) algorithm designs for Aptina SOC.• Lens shading and distortion correction, global and local tone mapping, nonlinear pixel processing, multiframe and interlaced HDR.

Image processing algorithm design including Bayer denoise, multi-exposure HDR, global/local tone mapping, lens shading and lens distortion correction.HDR algorithms in blending and tone mapping for automotive and security applications CMOS 4-T, 5-T, 6-T pixel design and dark current reduction by identification of defect types by activation energy evaluation