Samuel AbatnehResume ->

AI engineer · Addis Ababa, Ethiopia

Samuel Abatneh

I build reliable AI products and backend systems, from retrieval and inference pipelines to APIs that hold up under real-world load.

See selected workRead my Resume Email me

Independent work

Upwork ->
$30K+
total earnings
17
jobs completed
3,596
hours worked
Top Rated Plus on Upwork

Profile

My work sits where applied AI meets dependable software. I care about retrieval quality, latency, observability, and whether a system is genuinely useful once it reaches users.

Selected work

Products and systems I've built.

Case studies spanning AI search, agents, learning tools, and real-time inference.

01

Flagship RAG Product

BibleClass AI

A theological search and study platform designed to answer doctrinal questions with high-confidence retrieval, multilingual understanding, and production-ready backend orchestration.

What I delivered
  • Ingestion pipeline for scripture, commentary, and structured theological resources.
  • Embedding workflow with multilingual retrieval and relevance reranking.
  • Confidence routing to decide when to answer, clarify, or fall back to broader search.
02

Autonomous Agent Systems

OpenClaw

An AI coding agent concept focused on terminal execution, MCP-style tool calling, memory, VFS isolation, and autonomous planning for developer workflows.

What I delivered
  • Planner loop decomposes tasks into executable steps with tool selection.
  • Terminal execution and VFS access isolate file operations from higher-level planning.
  • Memory layer stores goals, observations, and recovery context between actions.
03

Content Automation

AI Course Builder

A pipeline for turning raw learning materials into structured AI-assisted courses through scraping, chunking, vector indexing, and document generation.

What I delivered
  • Scrapers collect content from source pages and normalize metadata.
  • Chunking and embeddings prepare material for retrieval and curriculum assembly.
  • Vector search powers concept lookup and lesson generation.
04

Real-Time Inference

AI Live Streamer

A real-time interactive AI stack for voice, avatars, and GPU-backed generation optimized for low-latency streaming experiences.

What I delivered
  • Text-to-speech and query generation feed a low-latency response loop.
  • RAG provides contextual grounding for conversational answers.
  • Streaming transport coordinates avatars, speech, and incremental response tokens.

Experience

Engineering, research, and teaching.

2024 - Present

Machine Learning Engineer

Upwork

Working with international clients to build production AI applications, LLM systems, RAG pipelines, backend services, cloud infrastructure, text-to-image systems, and speech synthesis products.

2023 - 2024

AI Researcher

Addis Ababa University

Developed an NLP threat detection system for live social and communication contexts, later adopted for practical use by the Ethiopian government.

2021 - 2023

Algorithms Instructor

Africa to Silicon Valley (A2SV)

Mentored students on algorithms, competitive programming, backend projects, and hackathons while contributing to product development and high-performance problem-solving culture.

View full experience ->

Competitive programming

Strong foundations in algorithms and problem solving.

I have solved more than 1,000 algorithmic problems and taught competitive programming, turning that practice into clearer reasoning about performance, edge cases, and system design.

LeetCodeSamuelabatnehView profile ->CodeforcesSBAGEView profile ->

Contact

Have an ambitious AI or backend problem?

I'm open to engineering roles, research collaborations, and focused product work.

samuel.neuralnet@gmail.comgithub.com/samiabatlinkedin.com/in/samiabat