Francis Oundo

As a Developer at Turing, I fine-tuned financial LLMs via RLHF, designing 100+ compliance-focused prompts that boosted trade execution accuracy by 25% on models, leading to a 15% rise in user retention. My automated test suites slashed model hallucination errors by 15%, enhancing trust in AI-driven signals.

Also as an Analytics Engineer with 7+ years in fintech and AI, I design systems where data drives measurable business outcomes. At LOLC Kenya, I built real-time Power BI dashboards (Python/SQL) tracking customer credit behavior, enabling 40% faster loan approval decisions. Using PySpark and Azure Data Factory, I automated ETL pipelines for 10M+ transactions, reducing reporting latency by 25% and improving risk assessments.

Earlier, at KCB Bank, my SQL-based BI reports identified untapped SME segments, driving a 20% increase in cross-sell conversions. I also migrated legacy systems to Azure Synapse (Python), cutting query times by 30% for real-time fraud detection. Certified in Microsoft Fabric (DP-600) and Azure (DP-900), I bridge cloud engineering with AI—like optimizing a credit-scoring ML models that reduced defaults by 20%.

I’m passionate about high-stakes AI applications—like my LLM work optimizing algorithms—and thrive in collaborative, remote-first environments. My career is defined by translating complex data into scalable solutions, whether training models for fintech compliance or enabling self-service BI.