Open to opportunities

Chirag Sindhu

@chiragsindhu

Message

AI & DevOps engineer specializing in scalable generative AI systems.

India

Message

What I'm looking for

I seek roles building scalable production generative-AI systems with strong MLOps, low-latency real-time requirements, collaborative teams, and opportunities to optimize models and cloud-native architectures.

I am an AI and DevOps engineer with 5+ years deploying production generative AI systems, specializing in LLMs, RAG pipelines, LoRA/QLoRA, prompt engineering, and real-time conversational platforms. I have architected cloud-native AI solutions across AWS, GCP, Azure, and self-hosted environments, optimized models with quantization and distributed training, and delivered high-availability platforms with sub-100ms latency.

My hands-on work spans multimodal pipelines, real-time vision and speech systems, AR/VR and Metaverse applications, and scalable MLOps infrastructures using Kubernetes, Docker, Redis, PostgreSQL, and CI/CD. I drive performance gains via DeepSpeed/FSDP, INT8/INT4 quantization, and production integrations of Whisper, Gemini, and other LLM/ASR/TTS technologies to meet demanding SLAs and large request volumes.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Pioneverse

Oct 2023 - Present (2 years 8 months)

Architected low-latency multimodal AI pipelines and event-driven backends, enabling real-time conversational AI handling 50K+ requests and sub-100ms latency for streaming avatars and telephony integrations.

Python PyTorch TensorRT WebRTC Redis Kubernetes Docker RAG Prompt Engineering

Research & Development Engineer

MuseTalk

Apr 2022 - Oct 2023 (1 year 6 months)

Engineered real-time vision and conversational AI pipelines using OpenCV, PyTorch, TensorRT and deployed multilingual, low-latency systems on Kubernetes achieving 99.9% uptime and zero-downtime deployments.

PyTorch OpenCV TensorRT Kubernetes Docker Redis PostgreSQL Whisper

Research Engineer

Cognitive Systems Lab

Jul 2021 - Dec 2021 (5 months)

Designed cloud BIM/CAD visualization and built multilingual speech AI (Whisper, Wav2Vec2, HiFi-GAN) with <8% WER and 4.2+ MOS, and implemented distributed training to cut training time by ~70%.

Whisper DeepSpeed Fsdp Distributed Training WebGL