Looking for a job

Chrispine Owuor

@chrispineowuor

Message

AI Developer at SaidaLink AI fine-tuning models and building agents for multi-step tasks.

Kenya

Message

At SaidaLink AI, I fine-tune models for specific tasks and build and test agents that follow multi-step workflows. I review model outputs and refine training data and prompts to improve their responses.

At Wekkitech Limited, I managed cloud servers and built Prometheus and Grafana dashboards that reduced incident detection time by 40%. I also developed OneSacco, a financial application for transaction tracking, dispute resolution, and loan processing, using Java, Spring Boot, PostgreSQL, and AWS.

Experience

Work history, roles, and key accomplishments

Current

AI Trainer

Current

SaidaLink AI

May 2026 - Present (5 months)

Fine-tuned AI models for specific tasks and use cases to improve response quality and accuracy. Built and tested AI agents designed to complete multi-step tasks and follow defined workflows.

Model Fine Tuning AI Training AI Output Evaluation Prompt Engineering Automation

Associate Cloud Engineer

Wekkitech Limited

Apr 2025 - Nov 2025 (7 months)

Managed Linux and Windows servers across AWS and Azure environments to support reliable production systems. Configured and troubleshot DNS, DHCP and RBAC policies for client infrastructure.

Linux Windows Server AWS Azure DNS DHCP RBAC Backup And Recovery Prometheus Grafana

Associate Backend Engineer

Wekkitech Limited

Jan 2025 - Mar 2025 (2 months)

Managed PostgreSQL databases and wrote SQL queries that improved transactional data reliability by 30%. Built Bash and PowerShell scripts to automate routine configuration work and reduce manual errors.

PostgreSQL SQL BASH PowerShell Automation Debugging Logs Deployment

Network Operations Center Engineer

Jamii Telecommunications Ltd

May 2023 - Aug 2023 (3 months)

Monitored enterprise network infrastructure using SolarWinds to maintain over 99% service availability. Diagnosed and resolved more than 100 hardware and connectivity incidents while maintaining a 95% SLA resolution rate.