Mehmet – Python, LangChain, LLM

Developed core infrastructure for an AI agent platform supporting financial forecasting and research workflows. Built the organizational knowledge retrieval layer that enabled AI agents to access and reason over information from repositories, documentation, collaboration tools, and other internal data sources.

• Owned the end-to-end implementation of an organizational context layer for AI agents supporting financial forecasting workflows;
• Designed and implemented ETL pipelines that ingested repositories, commits, configuration files, Slack, Notion, and Google Drive content;
• Built agent-facing retrieval APIs for semantic search, symbol search, regex search, file discovery, and file retrieval;
• Implemented an OCL agent that accepted tasks from other agents, queried the organizational context layer, and returned grounded answers from indexed company knowledge;
• Created evals to measure OCL data quality, retrieval coverage, and answer usefulness for agent workflows;
• Integrated PostgreSQL/Drizzle metadata storage, ChromaDB vector search, and S3-compatible artifact storage for indexed context;
• Implemented webhook-based source synchronization, pipeline orchestration, job status tracking, cancellation, and recovery flows;
• Developed admin and API surfaces for managing sources, source groups, users, connections, pipelines, and webhook events;
• Containerized Python ETL services for code, config, commit, and document ingestion;
• Wrote unit, integration, and end-to-end tests for API routes, pipeline logic, ETL services, and agent retrieval behavior.

Developed an AI-powered banking intelligence platform that combined a semantic knowledge engine, a real-time digital twin for operational simulation, and an Auto-Code Engine that translated optimization decisions into automated system changes. Designed and implemented the Auto-Code Engine, enabling workflow configuration, automation deployment, and full auditability.

• Designed and implemented a code generation framework based on Yeoman Generator to automatically create Spring Boot microservices and Angular frontends from structured definition files in Git;
• Built CI/CD pipelines using infrastructure as code for automated build and deployment of generated applications to AWS-hosted Kubernetes clusters;
• Integrated Keycloak for federated authentication and SSO across generated applications;
• Delivered a fully functional Client Onboarding platform auto-generated by the framework;
• Collaborated with the founding team to establish the foundation for an end-to-end application generation platform.

An enterprise contract management platform enabling automated analysis and semantic search across thousands of legal documents.

• Developed AI components for an enterprise Contract Management platform;
• Built an OCR and NLP pipeline to import and process bulk contracts from PDFs using Tesseract OCR and spaCy;
• Implemented a Retrieval-Augmented Generation system using LangChain, ChromaDB, and OpenAI GPT models to power contract summarization, redlining suggestions, and similarity search;
• Exposed FastAPI endpoints for contract analytics integration;
• Developed a React and TypeScript chat interface for natural language contract queries;
• Delivered a production-ready AI-assisted contract review platform.

A Java-based Enterprise Service Bus (ESB) and Messaging Queue platform for building scalable, high-performance enterprise integration solutions.

• Served as Chief Architect, driving key architectural decisions across the platform;
• Reviewed and approved major code contributions to ensure quality and architectural consistency;
• Established engineering standards and provided technical leadership to a globally distributed development team;
• Designed and optimized multithreaded, highly scalable, and highly available solutions for mission-critical enterprise workloads.