Open to opportunities

Rahul Sankhwar

@rahulsankhwar

Message

I’m a Machine Learning Engineer building production AI systems across speech, vision, text, and LLM workflows.

India

Message

What I'm looking for

I’m looking to build and own production ML platforms—especially low-latency, real-time inference for speech, multimodal AI, and LLM workflows—where I can lead teams, scale systems in AWS or on-prem, and ship to enterprise customers.

I’m a Builder / Machine Learning Engineer with 8+ years of experience delivering production-grade AI systems across text, speech, computer vision, and LLM-powered workflows. I focus on end-to-end ML platform ownership—designing models and the real-time systems that reliably serve them.

At Observe.ai, I led speech ML initiatives by architecting services, evaluation workflows, and in-house production-grade real-time streaming ASR. I owned on-premises enterprise deployments, scaled containerized inference workers with Docker and Kubernetes, and supported real-time inference plus post-call analytics with high-availability requirements.

Previously, as Co-founder & CTO at Dubdub.ai (acquired by Observe.ai), I defined the AI roadmap and built an end-to-end production ML platform for audio and video generation. I delivered scalable backend services and REST APIs on AWS, engineered content analysis and transformation pipelines, and built LLM-powered applications and agentic orchestration to automate script generation, content transformation, and video clipping.

Earlier at ShareChat and Samsung Research Institute, I shipped production multimodal deep learning for content tagging and recommendation, built on-device face landmark detection, and drove core computer-vision platform ownership (Bixby Vision Engine) in C++. Across roles, I’ve consistently combined technical leadership, system design, and execution—scaling teams and products that enterprise customers can depend on.

Experience

Work history, roles, and key accomplishments

Lead ML Engineer (Speech)

Observe.ai

Aug 2024 - Dec 2025 (1 year 4 months)

Architected evaluation workflows and deployed production-grade real-time streaming ASR services for enterprise customers, including post-call analytics. Led on-premises deployments and high-availability speech processing pipelines, scaling containerized inference workers with Docker and Kubernetes.

Real Time AI Inference Stream Processing On Premise Docker Kubernetes Model Evaluation Speech Processing

Co-founder & CTO (ML)

Dubdub.ai

Apr 2021 - Aug 2024 (3 years 4 months)

Defined and executed the AI roadmap, building an end-to-end production ML platform for audio and video generation with scalable inference on AWS. Delivered ML backend services and REST APIs, built speech and content analysis pipelines (ASR and speaker diarization), and developed LLM-powered agents for content transformation and video clipping, culminating in a successful acquisition.

Audio Generation REST APIs Speaker Diarization Audio Pipelines Vector Databases AI Agents

Data Scientist II

ShareChat

Dec 2019 - Jul 2020 (7 months)

Built multimodal deep learning models for content tagging and recommendation in large-scale production environments. Implemented large-scale content intelligence clustering/filtering and developed on-device real-time face landmark detection using C++ and TensorFlow Lite.

Content Tagging Recommendation Systems Clustering Filtering Algorithms Real Time On Device Vision Face Landmark Detection C Tensorflow Lite

Lead Engineer / Senior SWE

Samsung Research Institute

Jul 2016 - Nov 2019 (3 years 4 months)

Owned the Bixby Vision Engine in C++, leading architecture, releases, and code reviews while delivering on-device computer vision capabilities for flagship consumer devices. Built data collection and active-learning-driven annotation workflows, developed scene optimization models, and implemented C++ vision libraries for obstruction removal and video summarization.

Data Annotation Active Learning Object Detection Summarization