Open to opportunities

Sai Amrit

@saiamrit

Message

I build production computer vision and AI systems for biometrics, synthetic data, and medical imaging.

India

Message

What I'm looking for

I'm looking to build and evaluate AI systems spanning computer vision, biometrics, generative models, medical imaging, and brain-signal analysis, while continuing research that translates into real-world products.

I've shipped production AI systems at Tech5 India, including a RAG and LLM-based chatbot, face recognition technology, and 3D talking-face generation for internal and customer-facing use cases.

Across Tech5, HyperVerge, Mercedes-Benz Research and Development India, and IIIT Hyderabad, I've built biometric pipelines for face verification, anti-spoofing, person re-identification, iris generation, synthetic data, and multi-view reconstruction. My work includes a 3D face generation and rendering pipeline and on-device object detection for real-time drone applications.

I've also curated 9 million question-answer pairs across 21,000 chest X-rays for GIV-CXR and published research at IJCB, ICCV Workshop, IJCNN, and a CVPR 2026 Medical Reasoning Workshop.

Experience

Work history, roles, and key accomplishments

Current

Independent Meditation Researcher and Coach

Current

Self-Directed Research

Oct 2025 - Present (11 months)

Conducts independent research on meditation and mindfulness, studying cognitive, emotional, and neuroscientific correlates using EEG and AI-based analysis. Designs and leads meditation workshops and guided sessions focused on attention and emotional regulation.

EEG AI Meditation Research Mindfulness Workshop Design

AI Research Engineer

Tech5 India

May 2024 - Oct 2025 (1 year 5 months)

Shipped a RAG and LLM-based AI chatbot to production and deployed face recognition technology. Developed 3D talking face generation, iris generation methods, and integrated multi-view reconstruction and biometric representation learning modules into production pipelines.

RAG LLM Face Recognition 3D Face Generation Iris Generation Multi View Reconstruction Biometric Representation Learning

Research Intern

HyperVerge

Dec 2023 - May 2024 (5 months)

Contributed to a face recognition initiative using self-supervised pretraining, improving representation quality for production identity-verification models.

Face Recognition Self Supervised Learning Representation Learning

Research Intern

Mercedes-Benz Research and Development India

Jan 2023 - Dec 2023 (11 months)

Built a 3D face generation and rendering pipeline using SDF-based mesh generator, FLAME-based texture and lighting modules, and StyleGAN super-resolution to generate synthetic training data for downstream face models.

3D Face Generation SDF Flame StyleGAN Rendering Synthetic Data

Research Associate / Research Assistant

Biometrics and Secure Identity Lab (BASIL) and CVIT, IIIT Hy

May 2019 - Jun 2020 (1 year 1 month)

Built a face verification and classification module achieving over 98% accuracy and a face anti-spoofing module achieving 97% accuracy. Generated a large-scale synthetic Indian face dataset and deployed an object detection pipeline on NVIDIA Jetson Nano for drone applications.

Facial Recognition Antispoofing Pruning Quantization Synthetic Data Generation Object Detection NVIDIA Jetson

Education

Degrees, certifications, and relevant coursework

International Institute of Information Technology (IIIT-H)

M.S. by Research, Computer Science and Engineering

Grade: 8.75/10

M.S. by Research in Computer Science and Engineering, with thesis on robust and secure face presentation attack detection. Achieved a CGPA of 8.75/10.