At Asinnah, I built the Asinnah Academy e-learning app in Flutter, with anti-screen-capture protection and encryption. I added an AI study assistant that answers student questions from course documents.
I engineered its RAG pipeline using chunking, embeddings, and vector search with LangChain, OpenAI, and LiteLLM. I also ran SDLC security using SAST and DefectDojo ASPM, integrating my thesis’s LLM triage to cut noise by 80%.
At Julep AI, I built a social-media intelligence system that used hundreds of sub-agents to gather and filter content, research trends, and generate and publish AI images and videos after human review. I acted as product owner for seven enterprise clients and led development of the Memory Store dashboard.
At Q Deep, I built a quantum-optimization platform and a hybrid decomposition pipeline that cut solve time around 40% versus the classical solver. My work also includes an MCP-based agentic system for triaging security-scanner findings, which achieved 80–91% classification accuracy across seven models.
