Alimi David

• Architected and deployed the Hoydoon cross-platform mobile application using React Native, successfully launching on both the iOS App Store and Google Play Store to deliver a low-latency, full-stack user experience.
• Addressed the challenge of inefficient property discovery by developing a robust marketplace API with Node.js, TypeScript, and MongoDB, implementing advanced geolocation filtering that reduced average user search time by 40%.
• Eliminated payment friction for diverse global markets by engineering a multi-provider gateway strategy (Stripe, Paystack, Waafi, Sifalo) with automated webhook verification, reducing failed transactions and expanding regional market access.
• Prevented stalled negotiations and buyer drop-off by engineering a real-time chat architecture using Socket.io and message caching, which accelerated agent responsiveness and increased platform user retention.
• Automated CI/CD pipelines using Azure DevOps and GitHub Actions, reducing deployment time by 50%. Integrated Sentry for error tracking and utilized AI tools (Claude) to identify vulnerabilities, debug pipeline failures, and script automated deployment workflows, drastically slashing MTTR.

• Addressed the challenge of autonomous agent reliability by building a custom AI agent framework using LangGraph, which achieved a 76% GAIA benchmark score in internal evaluations.
• Solved the problem of isolated AI contexts by implementing secure third-party integrations (Gmail, Notion, Jira etc.) using Composio, empowering agents to autonomously access user data and execute complex cross-platform workflows without manual input.
• Solved the problem of siloed communication data by architecting a scalable email-to-task automation system; utilised IMAP/SMTP and Redis queues to enable real-time task creation directly from email inboxes.
• Improved system throughput and user experience by migrating the backend from synchronous to asynchronous database connections, effectively eliminating blocking I/O overhead.
• Accelerated the end-to-end delivery of agentic workflows by rapidly assembling responsive frontend interfaces, leveraging Claude Code and Google Antigravity to drastically reduce UI development time and connect users seamlessly to underlying AI agents.

• Resolved the critical issue of fragmented vendor data by architecting a centralised bid management system using Django and PostgreSQL, enabling real-time price comparisons that increased profit margin visibility by 20%.
• Eliminated manual data entry bottlenecks by engineering a customised Django Admin interface with dynamic auto-population and inline forms, which reduced average bid submission time by 40%.
• Solved the challenge of untraceable database operations by building thread-local middleware for automatic user context tracking, ensuring 100% audit trail compliance without manual intervention.
• Reduced production deployment risks by scripting automated database migrations and Gunicorn configurations, cutting deployment downtime to near-zero.

• Architected and deployed the core RESTful APIs for the official Zepama mobile application (available on the Google Play Store), ensuring high availability and seamless data sync.
• Optimised ORM queries and database structure, improving backend performance by 20%.
• Authored comprehensive API documentation, detailing endpoints, parameters, and usage guidelines to streamline third-party integration and reduce onboarding time.

Grade: First Class Honours (GPA 4.71/5.00)

Built an autonomous agent framework using LangGraph and Google Gemini with a FastAPI control plane for efficient bot lifecycle management and health monitoring. Engineered a RAG pipeline using PostgreSQL (pgvector), implementing real-time message embedding and background processing for semantic context retrieval. Stack: Python, LangGraph, FastAPI, PostgreSQL (pgvector).

Developed a full-stack collaborative application, rapidly assembling a responsive UI using React and Tailwind CSS, while utilizing Google Gemini to generate interactive mind maps directly from user prompts. Architected a real-time synchronisation engine with offline state management and connection recovery, ensuring zero data loss during network interruptions. Stack: React, Node.js, WebSocket, Gemini AI.