Headline

Applied Scientist II at Amazon Prime Video, Former Generative AI Tech Lead - Harman, Camera Modelling Research Engineer - Qualcomm India | Research Scholar at IIIT Hyderabad | Computer Vision | DL | ML | IoT

About

I am working as an applied scientist II at Amazon where I work with Prime Video Compression Effeciency Research team. Previously, I worked as a generative AI tech lead at Harman where I collaborate with IQVIA to develop healthcare solutions. Before that, I was working as a camera modelling engineer at Qualcomm multimedia team where I was responsible for managing end-to-end Image Signal Processor (ISP) pipelines. I also interned with the Amazon India Machine Learning team, collaborating with exceptional individuals on projects involving Graph Neural Networks. My research interests encompass various domains, including Computer Vision, Computational Photography, Generative AI, Environmental Sensing (Air Pollution Monitoring), Machine Learning, and the Internet of Things (IoT). I have completed MS by Research from IIIT Hyderabad where I was advised by Prof. Sachin Chaudhari. I broadly worked in the area of ML/DL based IoT to monitor the air pollution based on the traffic on roads. Our team successfully deployed this work on the road of Hyderabad to detect the Air Quality Index in real-time. Previously, I was advised by Prof. C.V. Jawahar where I worked on sports analysis focusing mainly on Badminton. The work included player localization (for both singles and doubles matches), shots played by players and their overall mobility (player's speed on the court, distance covered by players etc.). The framework was deployed on live games at Premier Badminton League (PBL) for real-time analysis. To know more about me, visit my webpage -https://nitinnilesh.github.io

Location

Bengaluru

AI stack

claude code

codex

gemini

github copilot

kiro

langchain

Work experience

Applied Scientist II at Amazon, Jan 2025 - Present · 1 yr 9 mos, Bengaluru, Applied scientist focused on video quality evaluation, encoding optimization, and production streaming efficiency. Built ML-based quality metrics, mobile-optimized encoding ladders, and agentic AI workflows to improve perceptual quality, reduce bitrate, and accelerate experimentation. Led the encoding research that enabled the unification of Amazon Prime Video and MX Player on mobile, designing a saliency-aware ladder spanning 1080p to 144p. Delivered ~25% bitrate savings, improved floor VMAF by ~20%, and reduced streaming consumption to nearly 0.2 GB/hour.

Generative AI Tech Lead at HARMAN International, Feb 2024 - Jan 2025 · 1 yr, Bengaluru, Worked with IQVIA for the development of a healthcare-domain GenAI/RAG chatbot using LLMs, embeddings, semantic retrieval, and FastAPI. Built document intelligence and symptom summarization workflows to convert hospital protocol documents into an interactive knowledge assistant.

Camera Algorithm Modeling Framework Engineer at Qualcomm, Sep 2021 - Nov 2023 · 2 yrs 3 mos, Bengaluru, At Qualcomm, my role entails overseeing the entire imaging pipeline, which spans from the initial sensor input to the production of the final image. This involves the implementation and management of crucial camera techniques like filtering, sharpening, denoising, compression, and other essential processes to enhance the overall image quality and optimize the imaging experience for end-users.

Mentor at TalentSprint, Sep 2018 - Sep 2021 · 3 yrs 1 mo, Hyderabad Area, India, I worked as a part-time mentor at AI-ML course offered by Talent Sprint in collaboration with IIIT Hyderabad. I taught machine learning and deep learning concepts to the industry professionals.

Research Scholar at IIIT Hyderabad, Jul 2017 - Jul 2023 · 6 yrs 1 mo, Hyderabad, Telangana, India

Education

Master of Science - MS, Computer Science, International Institute of Information Technology Hyderabad (IIITH)

Bachelor of Technology (B.Tech.), Computer Science Engineering, Institute of Engineering & Management, Kolkata

Public links

github: https://github.com/nitinnilesh

leetcode: https://leetcode.com/seedhaLadka

youtube.com: https://youtube.com/@nitinnilesh

nitinnilesh.github.io: https://nitinnilesh.github.io/