Kagan Demirhindi

Product experience platform that helps companies make product intelligence actionable at speed and scale.

• Agentic AI Development: Played a key role in building and shipping the Supervisor Agent for Pendo’s Agent Mode. Worked across teams to integrate sub-agents and improve response accuracy to production-ready quality.

• AI Workflow Automation: Designed and shipped custom agentic workflows for context-aware outputs, with a strong focus on reliable orchestration and response quality.

• Agent Evaluation & Observability: Built evaluation workflows using LLM-as-a-Judge and LangSmith to measure agent quality and speed up iteration.

• AI Enablement & Customer Feedback: Ran internal AI trainings, onboarded teams starting to work in AI and regularly joined customer calls to feed user context back into product and agent decisions.

The only project management platform built specifically for client work.

• AI Product Development: Worked on AI features from ideation to production. Shipped AI-generated project setups that doubled onboarding completion rates. Built AI Forecasting using TinyTimeMixer and Prophet, self-hosted on AWS.

• Full Stack Development: Built production features with Vue.js, Tailwind CSS and Go. Used LaunchDarkly feature flags for safe, incremental rollouts.

• Growth & User Analytics: Ran data-driven experiments to optimize KPIs and drive user growth. Used Pendo, HubSpot, Intercom and Hotjar to understand user behavior and measure impact.

AI-powered t-shirt design and print-on-demand web app. Built from scratch, shipped to production.

• Built an AI product from idea to deployed app. Vue, Vite, Nuxt 3 on Cloudflare Pages with SSR. Integrated Stable Diffusion API for designs, Stripe for payments, Cloudflare R2 for storage.

• Learned fast by shipping. The project validated technical feasibility but hit payment processing issues during customer validation. Demo still available at looksounique.com.

London-based fintech startup. ePOS app that turns tablets and smartphones into powerful cash registers.

• Small team, high autonomy. Worked directly with CEO and CTO to define the product. Learned to be outcome-focused and ship fast.

• Built Go microservices with gRPC. Dockerized and deployed on Google Kubernetes Engine.

• Integrated with partner bank API to automate merchant onboarding, reduced time from contract to first transaction to just hours.

• Shipped a React + TypeScript mobile app using Ionic and Capacitor. Built custom Capacitor plugins for native SDK access.

Software consultancy in Izmir, Turkey. Strong culture of learning and exploring new technologies.

• Shipped multiple projects as an intern: product catalog app, activity monitoring software and data pipelines. Learned by building, not just reading.

• Built microservices with Java + Micronaut, PostgreSQL, GraphQL. Containerized with Docker and Docker Swarm.

• Built web scrapers with Scrapy-Splash to extract gigabytes of data daily. Prepared data for ML training using Python.

• Data visualization and analysis with Pandas, Elasticsearch, Kibana. Automated deployments with Bash.

“How to Train Your (Coding) Agent”: Explored the current state of LLM capabilities along with practical patterns for managing context windows, subagents, skills and agentic workflows (bash loops) for AI-first development.

“Fewer Juniors Today = Fewer Seniors Tomorrow”: Sparked widespread industry discussion on the long-term talent pipeline risks of AI hype in software engineering.

Top research university in Turkey.

Started working as a developer in my second year while balancing coursework. Active in Ege Entrepreneurship Society, organizing company visits and campus events. Learned early that building is the best way to learn.