Staff Machine Learning Engineer
Intel Corporation · Folsom, CA, USA · May 2024 – Present
- Designed and implemented transformer-based and attention-based models for NLP and vision tasks on AI PCs and edge devices.
- Developed novel algorithm reducing LLM model size by over 75% — 7.5× beyond the original 10% goal.
- Pioneered a latency-reduction technique cutting LLM CPU inference by over 25% (goal was 5%).
- Built real-time sign language translation pipeline: video → ASL labels → quantized LLM → English text.
- Tech Lead for ISVs deploying Mixtral-8x7 and enabling AI features on AI-enabled PCs and edge devices.
- Mentored junior engineers on ML/software optimization, defined performance metrics and benchmarks.
OpenVINOLLM OptimizationINT4 Quant.Sign Language AIEdge DeploymentMixtral
Imaging Scientist II
Philips Pharma Solutions · Rochester, NY, USA · May 2021 – May 2024
- Developed CV and DL methods for 3D stitching, registration, segmentation, super-resolution, and out-painting.
- Built full-leg image reconstruction system to calculate knee angle — a first for Philips.
- Reduced dataset loading time from 2 hours to 10 minutes via optimized data pipeline design.
- Developed 3D MRI alignment algorithms to reduce cost of repeat imaging sessions.
- Designed advanced NLP models for text classification, sentiment analysis, and language generation.
- Tech Lead for MRI and CT imaging evaluation workflows.
3D ReconstructionMedical ImagingKnee Angle AIMRI / CTNLP / LLMsSegmentation
Medical Imaging Analytical Intern
Siemens Healthineers · Malvern, PA, USA · June – August 2019
- Developed a DL algorithm classifying echocardiogram views with 95% classification accuracy.
- Built image processing tools for dataset preparation and classification pipeline.
- Analyzed and proposed optimal deep learning strategies for the cardiac imaging task.
Echocardiogram AI95% AccuracyCardiac ImagingClassification
Research Assistant & Software Developer
University of Oklahoma · Norman, OK, USA · Jan 2017 – May 2021
- Published 13+ papers in top-tier journals across ML, deep learning, medical imaging, and data science.
- COVID-19 detection from chest X-rays with 94.5% accuracy across three classes.
- Hybrid CNN optimization (Xception, InceptionV3) for lesion classification: 91% accuracy.
- Breast cancer risk prediction: 9.7% improvement over baseline via bilateral radiologist-mimicking scheme.
- Designed speech segmentation algorithms and image steganalysis systems using ML.
13+ PublicationsCOVID-19 DetectionBreast Cancer AICNN OptimizationSteganography
