Tural Gulmammadov

Founder and CEO at Mergen AI

Applying Graphs + AI for Compliance Work

GitHub | LinkedIn | Twitter/X | Medium | Substack | Email

About me

I learned chess from my grandpa when I was 5 years old and won a couple of local chess tournaments during early elementary school. While training on USACO and ACM ICPC-style problems through national and international programming olympiads, I got exposed to graph theory, computational geometry, dynamic/greedy programming, and data structures early high school years. I started my programming journey with Pascal then C++ during those years.

Today, I’m the Founder and CEO of Mergen AI, where I design and build large-scale AI systems that execute regulatory workflows. Previously, I served as Head of Core Generative AI Development at Oracle, leading cross-functional teams of data scientists and AI/ML engineers responsible for production-grade generative and applied AI platforms deployed across multiple industries.

Professionally, I have been building AI/ML systems since 2015. Started my career building ETL jobs, AI/ML models, and online/offline inference APIs for various size of startups in Toronto, ON. I approach AI as an engineering discipline — focused on correctness, scalability, interpretability, and real-world constraints — bridging theoretical foundations with production systems.

Latest news

I spend most of my time validating the business case for Mergen AI and building Claude Code skills to solve my day-to-day development problems — like build-kg. Still doing creative work with uber-polya — a math problem-solver skill for AI coding assistants, defaulted on Claude Code. Both are open-source and used at Mergen AI’s compliance AI engine and workflow automations.

Projects

• First Proof? — Solutions to the 1stproof.org challenge: 10 research-level math problems solved autonomously by uber-polya
• uber-polya — The first math problem-solver for AI coding assistants
• build-kg — Turn any topic into a structured knowledge graph
• build-kg-viewer — Graph visualization for Apache AGE with AI-powered queries

View All →

Teaching

I designed and taught an AI/ML Operations course at a joint program between the Vector Institute and the University of Toronto — focused on the design, deployment, and operation of reliable machine learning systems, with emphasis on AI/ML Ops, distributed training, observability, and cost-aware infrastructure.

What excites me

My engineering expertise lies in AI/ML Ops — delivering reliable AI/ML systems on a cloud scale. My technical interests center on the application of graph theory, discrete mathematics, and probabilistic reasoning to machine learning systems operating in distributed and heterogeneous compute environments.

I am a technically driven leader — equally at home deep in code and architecture as I am shaping product strategy and leading teams. Questions like how to solve consensus in a network of nodes when encountering a node failure while computing neighborhood walks over a distributed graph entertain me the most. Equally — designing a generative AI roadmap that balances platform investment against near-term product delivery, while the underlying models shift faster than your release cycles, and keeping a team of scientists and engineers aligned on execution through it all.

Publications

A Practical Guide to Oracle AI Engineering — Author 2026

In A Practical Guide to Oracle AI Engineering, you’ll learn how to tackle the challenges of building scalable, high-performance AI workflows in modern enterprises. Many organizations struggle to turn raw data into actionable insights while maintaining security, compliance, and operational efficiency. This book provides practical, end-to-end guidance for data engineers and architects to design, secure, implement, and optimize ML and GenAI solutions across Oracle Cloud, Oracle Database, and MySQL HeatWave. Written by multiple Oracle experts with deep experience in Oracle technologies and enterprise data platforms, this book walks you through real-world examples and hands-on workflows, from data preparation and in-database ML to deploying GenAI-powered applications and intelligent agents. You’ll gain skills in building pipelines, managing models, leveraging vector search for advanced AI use cases, and integrating AI into business applications with APEX and Oracle Digital Assistant. Advanced topics include scalable model deployment, serverless inference, monitoring, and MLOps best practices. By the end, you’ll be equipped to solve complex data challenges, accelerate AI adoption, and deliver measurable business impact through intelligent, production-ready solutions.

🎟️ Get 20% off — apply code 20ORACLEAI at checkout on the Packt website.

Graph Machine Learning — Technical Reviewer April 2021 – June 2021, Toronto, ON

Graph Machine Learning introduces you to a set of tools used for processing network data and leveraging the power of the relationships between entities for predictive, modeling, and analytics tasks. The first chapters introduce graph theory and graph machine learning, as well as the scope of their potential use. You’ll then learn all you need to know about the main machine learning models for graph representation learning: their purpose, how they work, and how they can be implemented across a wide range of supervised and unsupervised learning applications. You’ll build a complete machine learning pipeline — including data processing, model training, and prediction — to exploit the full potential of graph data. After covering the basics, you’ll be taken through real-world scenarios such as extracting data from social networks, text analytics, and natural language processing.

Patents

Discrete Optimization of Configuration Attributes — USPTO 12,541,681 October 2020 – April 2021, Toronto, ON

A graph-based discrete optimization algorithm designed to optimize mobile plan offerings by customer segment for world-leading telecom providers.

Awards

Beyond work

Reading books in cognitive science, statistics, and psychology. I enjoy playing chess, painting, seasonal horse riding, and paddling. Spending most of my personal time with my daughter and wife in Toronto, ON.