Open to opportunities

Ricky Ajedev

@rickyajedev

Software engineer blending cybersecurity, high-performance C#, and applied machine learning for secure AI-ready systems.

Kenya

Message

What I'm looking for

I’m looking for a role where I build secure, high-performance software (C#/.NET) and contribute to AI training and evaluation—partnering with ML and security teams to ship measurable improvements in safety, alignment, and real-world reliability.

I’m a Jomo Kenyatta University of Agriculture & Technology (JKUAT) alumnus in Mathematics & Computer Science, with a rare mix of mathematical rigour and pragmatic engineering. My focus is building secure, scalable, production-grade systems while rigorously evaluating machine learning and AI behavior.

In AI Training, I’ve handled high-quality RLHF and SFT work on code generation, reasoning, and instruction-following tasks, evaluating accuracy, safety, and alignment across 5,000+ outputs. I also specialize in image generation and generative model evaluation, providing expert feedback against reference benchmarks.

On the engineering side, I’ve architected high-throughput backend services with C# / .NET 8, delivering sub-100 ms latency at 50k+ API requests per minute. I implement security-focused middleware (JWT, SHA256 request signing), anomaly detection pipelines, and event-driven processing using Apache Kafka to reduce ETL execution time by 37%.

I’m equally committed to security research and reverse engineering, authoring 25+ technical articles and working on topics like SHA256 analysis, malware reverse engineering, binary exploitation, and cryptographic implementation risks. I aim to keep improving systems and models through measurable results, careful evaluation, and collaborative learning with ML and security teams.

Experience

Work history, roles, and key accomplishments

Current

AI Training Expert (RLHF)

Current

Handshake AI

Jan 2024 - Present (2 years 5 months)

Performed high-quality RLHF and SFT annotation for code generation and instruction-following tasks across 5,000+ evaluations. Improved inter-annotator agreement by 18% and provided expert feedback on diffusion model outputs against reference benchmarks.

SFT Data Annotation Prompt Engineering Safety & Alignment Inter Annotator Agreement

Current

Senior SAS Developer

Current

KenTech Solutions Ltd

Jan 2024 - Present (2 years 5 months)

Architected and maintained high-throughput C#/.NET 8 backend services handling 50k+ API requests per minute with sub-100 ms latency. Implemented JWT and SHA256 request signing with anomaly detection, and used Kafka with MongoDB pipelines to reduce ETL execution time by 37%.

C++.NET 8+ASP.NET Apache Kafka MongoDB JWT Anomaly Detection ETL

Current

AI Training Expert (Prompt)

Current

Outlier AI

Jan 2023 - Present (3 years 5 months)

Authored and evaluated adversarial, edge-case, and domain-specific prompts to stress-test LLM capabilities in mathematics, code, and cybersecurity. Trained junior annotators on evaluation rubrics, increasing team throughput by 25%, and produced first-author write-ups on model failure modes used to improve training data.

Prompt Engineering Adversarial Testing LLM Benchmarking Code Evaluation Rubrics Model Comparison Technical Writing

Data Science Analyst

Savanna Analytics Kenya

Jan 2023 - Aug 2023 (7 months)

Analyzed 10+ years of soil composition and microclimate data to optimize water usage and delivered a predictive irrigation scheduling model that reduced water consumption by 18% while maintaining crop yield. Automated recurring Excel reporting with openpyxl and VBA macros and improved non-technical team productivity by 30%.

Python Pandas predictive modeling Excel Automation Openpyxl Power BI Financial Modeling Data Analysis

Machine Learning Engineer Intern

AgroSmart Kenya Ltd

Built an end-to-end computer vision pipeline using OpenCV and PyTorch to detect crop diseases from drone imagery, achieving 94% accuracy. Engineered time-series yield prediction models using weather and soil moisture data, improving forecast precision by 22%, and deployed models on AWS EC2 with Docker.

OpenCV PyTorch Computer Vision Time Series Models Docker AWS EC2 Python

Education

Degrees, certifications, and relevant coursework

Jomo Kenyatta University of Agriculture & Technology (JKUAT)

Bachelor of Science, Mathematics & Computer Science

2022 - 2026

Grade: GPA 3.96/4.0

Activities and societies: Dean's List (all semesters); JKUAT CS Fellowship recipient; relevant coursework includes Advanced Algorithms, Cryptography, Reverse Engineering, Machine Learning, Distributed Systems, and Real Analysis.

Pursuing a Bachelor of Science in Mathematics & Computer Science at JKUAT (2022–2026). Received a CS Fellowship and achieved a 3.96/4.0 GPA with Dean's List distinction.