Skip to content
View eandualem's full-sized avatar
:octocat:
Open-Source Enthusiast
:octocat:
Open-Source Enthusiast

Block or report eandualem

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
eandualem/README.md

Hi, I'm Elias Andualem 👋🏾

Lead Generative AI & Agentic Engineer. I build production AI systems: multi-agent orchestration, MCP infrastructure, RAG, and the LLM evaluation that keeps them honest, plus the full-stack platforms they run on. About five years in, most recently leading Generative AI at Arclio and Tenacious.

📍 Based in Addis Ababa, Ethiopia (UTC+3), working remotely with US teams. 🔭 Open to Lead / Senior Generative AI roles: fully remote (contract or EOR, no sponsorship needed), or on-site with relocation and sponsorship.

LinkedIn Email


What I build

  • Agentic systems. Multi-agent orchestration, tool use, and workflow engines (LangGraph, Pydantic AI, Claude Agent SDK, XState).
  • MCP infrastructure. Multi-tenant hubs and the servers that plug into them (Google Workspace, Weaviate, S3).
  • RAG and evaluation. Hybrid retrieval on Weaviate, with RAGAS and QASPER harnesses so I can show retrieval actually got better, not just claim it did.
  • The platform underneath. FastAPI services and multi-provider LLM stacks (Anthropic, OpenAI, Gemini, Mistral), with NIST AI RMF on the governance side.

Recent roles

Arclio · Lead Generative AI Engineer · Apr 2025 – Mar 2026 Built a "Vercel for MCP" platform: a multi-tenant MCP hub, a no-code engine that turns plain-English intent into agent runs, hybrid (BM25 + vector) RAG on Weaviate, and Janus, a sandbox that runs untrusted tool calls without trusting them.

Tenacious · Lead Generative AI Engineer · Oct 2023 – Apr 2025 Shipped multi-channel AI assistants in production, LangGraph extraction agents that pushed data accuracy up about 70%, and evaluation pipelines that lifted reliability around 50%.

Fuller history on LinkedIn.

Tech

Languages

Python  TypeScript  JavaScript 

AI / Agents

LangChain  LangGraph  MCP  Pydantic  OpenAI  Anthropic 

Backend & Web

FastAPI  Node.js  React  Vue  Nuxt 

Data & Infra

PostgreSQL  Redis  Weaviate  Docker  AWS  GCP 


LinkedIn Email

Pinned Loading

  1. agent-backbone agent-backbone Public

    A lightweight control plane for terminal coding agents. Message, manage and delegate across Claude Code, Codex and OpenCode sessions, with GitHub Issues as the task list. No files in your repo.

    Python 4 1

  2. assistant-runtime assistant-runtime Public

    Put an assistant inside your own application: streamed turns over Socket.IO or AG-UI, sessions as message trees, tools your app executes, and prompt artifacts the assistant can revise. Built on Pyd…

    Python

  3. design-studio design-studio Public

    A Markdown and Mermaid design-document editor with an assistant panel, demonstrating assistant-runtime

    TypeScript

  4. xstate-mcp xstate-mcp Public

    An MCP server that gives AI coding agents live read/write access to running XState v5 state machines: inspect actors, query state and context, check transition eligibility, and send events.

    TypeScript 2 1

  5. Arclio/arclio-mcp-tooling Arclio/arclio-mcp-tooling Public

    Jupyter Notebook 3 3