Rohit Kumar

AI Researcher

Trivandrum, Kerala, India2 yrs 1 mo experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in Generative AI and Agentic Systems.
Pioneered lifestyle image generation pipelines.
Invited reviewer for top AI conferences.

Stackforce AI infers this person is a SaaS expert specializing in Generative AI and Computer Vision technologies.

Contact

contact@rohit.vision LinkedIn

Skills

Core Skills

Generative AiLarge Language Models (llm)Computer Vision

Other Skills

LangGraphRetrieval-Augmented Generation (RAG)FastAPIDockerQdrantDiffusion ModelsFluxControlNetPyTorchTransformer ModelsFlow MatchingStable DiffusionAgentic AI DevelopmentModel TrainingMachine Learning Algorithms

About

I build at the intersection of Generative AI and Agentic Systems, transforming research breakthroughs into production-grade software.Currently, as an AI Engineer at Armada AI, I architect end-to-end agentic workflows using LangGraph and RAG pipelines. My work focuses on enabling multi-step reasoning and autonomous decision-making for complex enterprise applications, backed by scalable infrastructure (FastAPI, Qdrant, Docker). Previously at Avataar, I pioneered lifestyle image generation pipelines using Diffusion Models (Flux, ControlNet), helping bridge the gap between 3D and 2D visual content. Core Competencies: - Agentic AI & LLMs: RAG Systems, LangGraph, Semantic Retrieval, Prompt Engineering. - Computer Vision: Diffusion Models, Object Detection (YOLO), Segmentation (SAM). - Research & Community: M.Tech in AI from IISc Bangalore (GATE AIR 221). Published at WACV 2025. Invited Technical Reviewer for CVPR 2026 and ECCV 2026.I am driven by the challenge of making AI reliable, scalable, and impactful. application, all while striving to advance the frontier of computer vision technology. Portfolio: https://rohit.vision

Experience

2 yrs 1 mo

Total Experience

8 mos

Average Tenure

1 yr

Current Experience

Armada

AI Engineer

Jun 2025 – Present · 1 yr · Trivandrum · On-site

Architecting autonomous enterprise agents and production-grade generative systems.
Agentic Workflows: Designing multi-step reasoning systems using LangGraph to automate complex decision-making processes.
RAG: Built high-performance Retrieval-Augmented Generation pipelines integrated with Qdrant for semantic search and context-aware responses.
Scalable Infrastructure: Developing robust ML APIs using FastAPI and Docker, ensuring low-latency inference for end-users.
Tech Stack: Python, Large Language Models (LLMs), LangGraph, PostgreSQL, Qdrant, Docker, FastAPI.

Generative AILarge Language Models (LLM)LangGraphRetrieval-Augmented Generation (RAG)FastAPIDocker

Avataar

Research Engineer

Jul 2024 – Apr 2025 · 9 mos · Bengaluru · On-site

Spearheaded the development of generative media pipelines, bridging the gap between 3D assets and photorealistic 2D imagery.
Generative Pipeline: Built an end-to-end lifestyle image generation system using Flux and ControlNets, enabling high-fidelity brand asset creation.
Diffusion Research: Modified diffusion sampling techniques and implemented intrinsic decomposition to significantly improve object reconstruction and visual consistency.
Vision Systems: Deployed zero-shot segmentation and classification models (combining SAM, YOLO-World, and CLIP) for automated asset understanding in low-data scenarios.
Tech Stack: PyTorch, Diffusion Models, Flux, ControlNet, CLIP, Qwen2.5, SAM.

Computer VisionGenerative AIDiffusion ModelsFluxControlNetPyTorch