h
hadeedsystems

Hadeed H

@hadeedsystems

AIFocused Full Stack Engineer

Pakistan
Engels
Sommige informatie wordt in het Engels weergegeven.
Over mij
I am a full stack engineer specializing in agentic AI systems, designing multi-agent architectures with LangGraph and CrewAI. I build production RAG pipelines and scalable Node.js and React applications end-to-end, from API design through deployment, with a focus on engineering intelligence.... Lees meer

Skills

h
hadeedsystems
Hadeed H
offline • 
Gemiddelde reactietijd: 1 uur

Bekijk mijn diensten

Full stack webapplicaties
I will build custom ai agents using langgraph or crewai

Portfolio

Werkervaring

Code_Maximus

Agentic AI Engineer

Code Maximus • Freelance

Oct 2026 - Nov 2026 • 1 mo

Designed and built a production-style conversational backend that qualifies real estate leads over WhatsApp in English and Arabic. Implemented secure Meta Cloud API webhooks with HMAC-SHA256 verification and two-layer idempotency, and a LangGraph agent that extracts budget, intent and timeline and books consultation calls. Delivered a paginated CRM admin API with 135 automated tests and an end-to-end security validation suite.

Softly_Wrapped

AI Engineer

Softly Wrapped • Fulltime

Mar 2026 - Jun 2026 • 3 mos

- Designing and building Stackmind, a multi-agent engineering intelligence platform that layers AI agents on top of a team's existing GitHub, Jira, and Confluence workflow to surface risk and blockers before they turn into fires. – Architected three purpose-built agents on the platform: a PR Risk Agent (LangGraph) that flags risky pull requests before merge, a Sprint Blocker Agent (CrewAI) that identifies what's actually slowing a sprint down, and a Postmortem Agent (LangGraph + RAG) that turns past incidents into searchable, actionable retrospectives. – Built the agent service as a standalone FastAPI/Python layer running alongside a Node/Express/MongoDB/Kafka backend and React frontend, so agents operate independently while staying in sync with the core platform.– Traced a silent RAG retrieval failure down to a missing vector search index and a trailing-space environment variable, then rebuilt the embedding pipeline to populate vector storage with 768 dimensional embeddings using Gemini's text-embedding-004 model. – Ported the platform's AES-256-GCM credential decryption logic from Node to Python so the agent layer could securely access each organization's encrypted GitHub and Jira tokens.