Building intelligence through generative AI & data engineering.
Anything about my work, answered from my real projects, with sources. (A live demo of what I build.)
Selected work.
Systems I took from messy data all the way to something people actually use.
Intelligent Form System
An agentic AI built on MCP servers that replaces manual form-filling: it ingests context from a document, a link, or plain text, asks follow-up questions to fill any gaps, then completes and submits the form for you.
Automated BI Reporting Pipeline
Replaced a company's slow, manual Excel reporting (7.2M+ rows, 1TB+ of data): modeled the data with dbt into an isolated reporting layer, built the client-facing dashboards in Amazon QuickSight, and automated a daily refresh that turns days of work into always-current reports.
Résumé Tailoring & Job Matching
Part of a platform that matches trainees to jobs: it generates a résumé tailored to each role by combining the candidate's skills with the job's requirements, then scores how closely they fit.
Contract Q&A RAG→
An optimized, RAGAS-evaluated retrieval pipeline that answers precise questions about legal contracts, built and measured for accuracy, not vibes.
Drone-Trajectory Data Pipeline→
A fully dockerized data-warehouse stack for traffic-trajectory data from swarm drones (pNEUMA): Airflow ingests large CSVs (~87MB each), and dbt builds tested, documented staging/production models for spatial-temporal analysis.
Amharic RAG Ad Builder→
An Amharic RAG pipeline that generates contextually relevant text ads for Telegram channels using open-source LLMs, bringing generative AI to a low-resource language.
Redash natural-language chatbot→
A Redash add-on that turns plain-English questions into SQL, so anyone can explore and visualize data without writing a query.
I'm an AI & Data Engineering professional in Addis Ababa. I build scalable data pipelines with Airflow and dbt, ship API-driven backends, and develop applied LLM systems like retrieval and document-processing pipelines, and I teach it all, training professionals at 10 Academy.
What I do
End-to-end AI and data engineering: scalable ELT pipelines (Airflow, dbt), API-driven backends, and applied LLM systems like RAG and document processing, with the evaluation to keep them honest.
How I work
I prototype quickly, validate with real data, and build things that hold up in production.
Experience
- AI Engineer & Technical TutorMar 2023 - Present10 Academy · Remote
Lecture and mentor across machine learning, data science, and data engineering, helping train and graduate 750+ practitioners (about half of them women), many of whom went on to a raise or a new, higher-paying role. Also build the data pipelines and APIs behind it all (incl. a résumé-generation platform): Airflow, dbt, PostgreSQL, and cloud analytics.
Education
- M.Sc. in Mechanical EngineeringDec 2021 - Jul 2024Addis Ababa University
- B.Sc. in Mechanical EngineeringDec 2014 - Jul 2019Dire Dawa University
Toolkit.
The stack I reach for, grouped by where it lives in the pipeline.
GenAI / LLMs
- RAG pipelines
- RAGAS evaluation
- OpenAI API
- Google Gemini
- Hugging Face
- Vector databases
ML / DL
- Machine learning
- scikit-learn
- NLP
- Low-resource (Amharic) NLP
Data / MLOps
- dbt
- QuickSight
- Airflow
- DVC
- Docker
- GitHub Actions
- PostgreSQL
Engineering
- Python
- FastAPI
- Flask
- Next.js / React
- SQL
- Git
What others said.
A little validation from people I've built with. (Samples, yours go here.)
“Kerod pairs serious LLM depth with a rare instinct for shipping. He took our prototype to a reliable, evaluated system far faster than we expected.”
Future Manager
Engineering Lead • A company you'll work with
“He bridges data, models, and product effortlessly, and explains results so non-experts can actually act on them.”
Future Collaborator
Product Manager • Another great team
One résumé, tailored to the role.
Pick a focus (LLM, ML, or Data Science) and the résumé re-selects, re-summarizes, and exports a matching PDF.
Open the CV builder →