Victor Adedeji

Architected and trained multimodal deep-learning systems that ingest, process, and align disparate high-dimensional datasets. Engineered self-supervised data pipelines lifting predictive accuracy on region-detection and pattern-recognition benchmarks.

Owned the AI backend infrastructure for an LLM-driven fintech product. Engineered real-time LLM streaming infrastructure (sub-100ms response latency) and built a production-grade agentic payment orchestration engine with human-in-the-loop guardrails.

Delivered a multimodal clinical-decision-support assistant — architected and integrated an end-to-end multimodal LLM-powered chatbot for real-time symptom analysis, drug interaction verification, and clinical decision support.

Lead AI backend engineer productionising generative-AI features and autonomous agents integrated into FastAPI microservices. Designed asynchronous messaging with RabbitMQ for decoupled inter-service communication across generative-AI services.