About Experience Skills Projects Research Blogs Contact CV ↓
Senior Manager, GenAI & Digital Health AI

Prateek
Singh

Deploying AI from research to real-world healthcare — at Samsung scale.

10+
Years in AI/ML
20+
Publications
3
Pending Patents
H100
GPU Research
Open to AI Leadership Roles  ·  Healthcare AI  ·  GenAI Systems
Scroll

Building AI that
lives on devices

I hold a PhD in Signal Processing and AI from IIT Roorkee and currently lead AI initiatives as Senior Manager of GenAI & Digital Health AI at Samsung Research Institute, Noida. My work sits at the intersection of on-device LLM deployment, physiological signal processing, and production-grade AI systems.

I've contributed to FDA-ready cardiovascular monitoring solutions, built intelligent RAG systems for clinical analytics, and shipped LLM features on constrained hardware. My research explores quantization, model compression, and efficient inference across CPUs, NPUs, and Qualcomm QNN.

Previously at Mercedes-Benz R&D India (in-cabin AI, 2022–2025) and Elastic Care, Canada (FDA-cleared AI modules, 2019–2022). Invited speaker at IIM Sambalpur, NIT Jalandhar, and other institutions.

On-Device AI LLM Quantization Digital Health ECG / PPG Processing FDA / SaMD GenAI Systems Qualcomm QNN RAG Architecture

Work Experience

From biomedical signal processing to production GenAI — a decade of building at the edge of AI and healthcare.

2025 — Present Current
Samsung Research Institute, Noida
Senior Manager — GenAI & Digital Health AI, Noida

Leading on-device AI initiatives across LLM deployment, GenAI systems, and physiological signal processing (ECG, PPG, audio) — targeting deployment across Samsung's 500M+ active device ecosystem. Built intelligent RAG systems for clinical data analytics, enabling clinicians to query physiological datasets in natural language. Driving NPU acceleration via Qualcomm QNN SDK and custom AI agent frameworks at production scale.

2022 — 2025 Automotive AI
Mercedes-Benz R&D India
AI Engineer — In-Cabin Intelligent Features

Developed in-cabin AI features for smart vehicles, applying multimodal sensing and real-time signal processing for driver monitoring systems. Two pending patents in automotive health sensing.

2019 — 2022 HealthTech
Elastic Care, Canada
AI Lead — FDA-Cleared Algorithms

Led development of FDA-cleared AI modules for wearable vital sign monitoring (HR, RR, arrhythmia classification). Supported FDA 510(k) submissions and SaMD compliance frameworks for cardiovascular and metabolic monitoring pipelines.

Ongoing Consulting
Independent Consultant
Regulatory AI & Health-Tech Advisory

Consulting for global health-tech startups including Vigo Health Care, Pareto Tree, and others on FDA 510(k), SaMD compliance, and clinical trial design. Invited speaker at IIM Sambalpur, NIT Jalandhar, and invited reviewer for IEEE TIM, IEEE IoT Journal, and Scientific Reports (Nature).

Skills & Stack

A decade of sharpening tools across signal processing, deep learning, and production AI systems.

My technical depth spans the full pipeline — from raw physiological signal acquisition to quantized model deployment on edge hardware. I work comfortably across research prototypes and production systems.

AI / ML Frameworks
PyTorch TensorFlow Transformers (HF) scikit-learn Keras
LLM / Inference
GGUF / llama.cpp TensorRT-LLM QNN SDK GPTQ / AWQ ONNX Runtime AIMET
Signal Processing
ECG / PPG HRV Analysis Wearable Sensors MATLAB SciPy
Infrastructure
Python Kotlin Android NDK Flask H100 GPU Docker
  • Python & Signal Processing95%
  • PyTorch / TensorFlow92%
  • MATLAB90%
  • LLM Quantization & Inference88%
  • On-Device AI (QNN / NPU)82%
  • Android / Kotlin + NDK75%
  • FDA / SaMD Compliance85%

Publications & Patents

Peer-reviewed contributions in deep learning, biomedical signal processing, and explainable AI — across Q1 journals and international conferences.

6 more publications across IEEE, Springer, and international conferences. View All on Google Scholar
Patents
View All on Google Scholar

Latest Blogs

Technical deep-dives into LLMs, quantization, inference runtimes, and digital health AI.

View All Blogs

Let's Collaborate

Open to AI leadership roles, research collaborations, consulting, and speaking invitations.