Software Engineer / AI Engineer
Yohannes Gezachew—
Building innovative software and AI solutions. Together we will create impactful digital experiences. Clean code, cutting-edge technologies.
The combination of my passion for algorithms, AI/ML, and user experience positions me in a unique place in the tech world.
Recent work
About
I'm Yohannes Gezachew Mokonnen, a Software Engineer and AI Engineer based in Ethiopia. With a strong foundation in Computer Science from Adama Science and Technology University and advanced training from A2SV's Competitive Programming Academy, I specialize in building innovative software and AI-driven solutions that combine technical excellence with user-centered design.
My experience spans from LLM training at Meta to leading full-stack development at Robotic Games, with deep expertise in backend engineering, RAG pipelines, and machine learning. I've built AI-powered applications like LissanAI, worked on healthcare platforms, and tackled NASA visualization challenges — always striving for clean code and impactful digital experiences.
When I'm not coding, you can find me solving competitive programming challenges on Codeforces and LeetCode, with 500+ problems solved. I'm always open to collaborating on exciting projects or discussing new opportunities.
Skills
Languages
Frameworks & Libraries
ML / AI
Databases
Cloud & DevOps
Experience
Designed and evaluated step-by-step LLM agent workflows across Windows, macOS, and Linux, identifying tool-use and reasoning failures to improve accuracy and safety. Collaborated with AI researchers and engineers to deliver ~25 reproducible LLM workflows per week, supporting accurate and reliable real-world task evaluations.
Directed development of an AI-powered platform that filters unwanted content and personalizes feeds across YouTube, X (Twitter), LinkedIn, and Reddit. Designed a dual-way local blocking algorithm and hybrid NLP + behavioral AI engine achieving 85%+ accuracy. Engineered a low-latency backend (< 200 ms) with FastAPI and Redis. Managed a 5-member cross-functional team, reducing feature rollout time by 80%.
Led a 5-person team to build LissanAI, an AI Amharic-to-English learning app for Ethiopians. Built backend with Go (Gin) and RAG pipelines for personalized, context-aware lessons. Implemented secure APIs, Google SSO, and Clean Architecture for scalability.
Built a Figma healthcare app prototype reducing booking time by 90% and improving financial transparency. Defined 5 MVP features and recommended a scalable React.js, Node.js, PostgreSQL stack to support 5x growth, delivering the MVP in 2 months.
Deployed 15+ on-site servers, Active Directory, and VMware Horizon for EPSS and ERA, achieving 99% infrastructure uptime and reducing downtime by 30%. Configured secure AD and VDI systems, improving accessibility by 60%. Conducted knowledge transfer sessions for over 25 clients.
Education
Relevant coursework: Data Structures & Algorithms, Design & Analysis of Algorithms, Information Retrieval, Artificial Intelligence, Computer Networks, Computer Architecture, Operating Systems, Concurrency, Object-Oriented Design, Advanced Java, C++ Programming, Requirements Engineering.
Advanced DSA training and software engineering. Solved 500+ problems across LeetCode and Codeforces.
Developed end-to-end ML pipelines; fine-tuned models in PyTorch; covered AI ethics, climate policy, land use & energy.
Awards & Certifications
Achieved 1st place in Ethiopia and 55th out of more than a thousand participants in the Middle East and Africa region.
Amazon Web Services foundational cloud certification.
Regression and Classification — Stanford University.