Open to opportunities

Afham Ahmad

@afhamahmad

AI/ML engineer building agentic LLM systems, RAG pipelines, and real-time computer-vision inference for industrial and robotics teams.

India

Message

What I'm looking for

I’m looking for a role where I can ship agentic AI and real-time computer-vision systems—working end-to-end from model/pipeline design to production integration with robotics or industrial data streams.

I build agentic AI systems with LangGraph, MCP, and ReAct—routing multi-agent LLM pipelines with RAG-based retrieval and tool execution tied to live GitHub REST APIs.

On the computer-vision side, I’ve shipped production OCR for industrial inspection, migrating a vision stack from Python to native C++ to cut latency and deliver real-time recognition on live line-scan camera feeds.

I also help lead robotics and perception efforts at Team Vyadh, working on low-latency video pipelines and a PyQt ground control station with ROS2 and MQTT control through multi-terrain field tests.

Experience

Work history, roles, and key accomplishments

Current

Machine Vision Intern

Current

KlarSYS Automations LLP

May 2026 - Present (2 months)

Built and evaluated a production CRAFT + PARSeq deep-learning OCR pipeline for industrial inspection, achieving 91% recognition accuracy on live line-scan camera feeds. Migrated the vision stack from Python to C++, reducing inspection latency 7x via cuDNN-accelerated fp16 inference.

Product Craft Parseq OpenCV CUDNN C++Python Socket Programming

Current

Technical Projects Lead / Vice Captain

Current

Team Vyadh, SEDS VIT

Sep 2025 - Present (10 months)

Coordinated execution across 2 competition teams (50+ members) spanning perception, autonomy, networking, and control, contributing to a 13th-place global finish at IRC 2025. Architected a GStreamer-based low-latency video pipeline, cutting camera feed latency from 5s to under 20ms.

GStreamer MQTT UDP TCP Team Leadership