At Careem, I develop enterprise AI agent platform capabilities for configuration, runtime execution, tools, connectors, knowledge resources, governance, and evaluation. I delivered a GitHub integration that lets repository events trigger AI agents and return AI-generated pull-request checks and comments.
I’ve extended agent execution with structured outputs and built a Golang scheduling service for recurring agent runs. I’ve also worked on secure streaming across deployment planes and collaborative agent-skill management.
At Wamo Labs, I designed and deployed agentic LLM orchestration systems using LangChain and LangGraph, integrating Azure OpenAI, AWS AI Services, and Google Vertex AI. I built transcription and multilingual video translation pipelines that processed more than 25,000 minutes of video monthly.
Earlier, at Devsinc, I worked on high-throughput backends, asynchronous task pipelines, and AI-powered products. At Turing, I supported model training and evaluation through annotation tooling, dataset curation, and human-in-the-loop testing; on the VendiGo project, I developed image-data workflows that contributed to YOLOv8 detection accuracy above 92%.
