resume

Senior AI/ML Engineer

ForesightCares Inc.

• Architected multi-agent AI workflows that automate complex healthcare compliance research and generate structured, review-ready documentation
• Built a cloud-native serverless AI service on AWS Lambda to deliver personalized exercise recommendations from curated clinical protocols
• Implemented scalable state-machine architecture for 5+ assessment protocols with automated monitoring and robust handling of multi-user scenarios
• Developed Field-of-View calibration model to correct 3D pose estimation across varying camera angles, improving measurement accuracy for at-home and clinical deployment scenarios
• Created 3D motion capture dataset using IMU-based mocap system for training and validation, collecting synchronized 60fps video and 400Hz motion data from 100+ subjects to train, validate and test our in-house pose estimator
• Designed a multi-stage signal processing pipeline that improved measurement stability and accuracy for joint-angle computation and gait analysis

Research Scientist

UNC Charlotte

• Developed a self-supervised foundation model with tokenized Vision Transformers, achieving 89.4% action classification and 76.56% AUC-ROC on anomaly detection; accepted to BMVC 2026

ML/AI Engineer

ForesightCares Inc.

• Architected and deployed an end-to-end AI/ML inference pipeline achieving 30fps real-time 3D human pose estimation, from data collection through production deployment and monitoring
• Engineered a cross-platform edge inference system with model optimization and deployment automation, eliminating cloud dependencies and reducing model bundle size through native infrastructure
• Co-invented AI-powered mobility assessment system (US Patent 12,343,138 B2), enabling automated fall risk evaluation and real-time exercise coaching via 3D skeleton modeling from monocular video

Research Scientist

UNC Charlotte

• Architected enterprise-scale distributed AI infrastructure enabling real-time multi-camera inference across hybrid edge/cloud deployments with privacy-aware data governance

Graduate Research Assistant

UNC Charlotte

• Designed an algorithm-architecture co-design framework for neural networks on heterogeneous compute platforms, enabling power-efficient inference through hardware-software optimization
• Developed a scalable real-time monitoring and reliability-prediction system for IoT infrastructure operating within a 1.87W power budget
• Created lightweight temporal convolutional network for vehicle trajectory prediction in highways, matching state-of-the-art accuracy with significantly reduced computational complexity (published in IEEE TITS, 97 citations)
• Co-developed an LLVM-based system architecture simulator for accelerator modeling (74 citations, IEEE/ACM MICRO 2020)
• Published 12+ peer-reviewed papers across IEEE and ACM venues with 657 citations (h-index: 11) spanning AI systems, model optimization, and distributed computing
• Designed attention-based temporal convolutional networks (ATCN) enabling fast time-series classification on resource-constrained embedded devices with formalized hyperparameters for application-specific architecture tuning (published in ACM TECS)
• Deployed ATCN on Cortex-M7 microcontroller consuming only 49% of 320KB RAM and 15% of 1MB flash; validated across 70 UCR 2018 time-series benchmarks using data augmentation (jittering, magnitude warping, window warping, scaling)
• Developed real-time person re-identification system at the edge using mixed precision (FP16/FP32) inference, achieving efficient tracking on embedded GPUs with MobileNetV2 backbone (published at ICIAR 2019, 17 citations)

FPGA Deployment Systems Engineer (Internship)

Astranis Space Technologies

• Supported FPGA deployment engineering for satellite systems, applying reconfigurable-computing and hardware/software integration expertise in an industry production environment

PhD

University of North Carolina at Charlotte

Electrical and Computer Engineering

• Advisor: Prof. Hamed Tabkhi
• Focus: Computer Architecture and Deep Learning

US Patent 12,343,138 B2: AI Powered Mobility Assessment System

ForesightCares Inc.

Inventors: H. Tabkhivayghan, M. Azarbayjani, R. Baharani

British Machine Vision Conference (BMVC), 2026 (accepted); arXiv:2502.05432

Self-supervised, LLM-inspired training for generalizable human motion understanding

IEEE IoT Journal, 2023

Ancilia: Scalable Intelligent Video Surveillance for the AIoT (89 citations)

ACM Trans. Embedded Computing Systems (TECS), 2022

ATCN: Resource-Efficient Processing of Time Series on Edge (6 citations)

IEEE Trans. Intelligent Transportation Systems, 2022

DeepTrack: Lightweight Deep Learning for Vehicle Trajectory Prediction in Highways (97 citations)

IEEE IoT Journal, 2019

Real-Time Deep Learning at the Edge for Scalable Reliability Modeling of Si-MOSFET Power Electronics Converters (74 citations)

IEEE IoT Journal, 2019

REVAMP2T: Real-Time Edge Video Analytics for Multicamera Privacy-Aware Pedestrian Tracking (93 citations)