Kagan Demirhindi

More than 14,000 companies use Pendo to create better software experiences, including Salesforce, Cisco, Verizon, HubSpot and Zendesk.

• Took Pendo's AI agent from pre-launch to production, building multi-agent orchestration, skills, prompt and context engineering, tool calling, MCP tooling, knowledge retrieval and semantic / vector search. Public adopters include Datasite and Meevo.

• Proposed and built Pendo’s AI evaluation tool (Python, React, PostgreSQL): LLM-as-a-Judge verdicts, repeatable test suites, repeated runs with pass thresholds for non-determinism, and comparison across agent harnesses including MCP and LangGraph setups. Engineers, PMs and managers now use it as an evaluation gate for model, MCP and launch decisions.

• Built tools for Pendo's MCP server, which lets customers reach their own product analytics from Claude and other AI clients. Adopted by customers including Cohley and Clinical Ink.

• Expanded the agent into Japan, from raising the opportunity to owning the regional deployment architecture and delivery required under data residency rules, writing the design and running it end to end with Legal and Platform Ops.

• Cut Pendo's agent error and issue rates in three months by debugging production traces and user feedback, and adding guardrails.

• Reduced unsupported requests to Pendo’s agent by shipping a knowledge-base retrieval (RAG) tool, after production traces showed that customers’ how-to questions were going unanswered.

• Made the agent safe to trust with enterprise data: redaction before retrieval reaches the model, every tool call scoped to the caller’s permissions so one customer’s data cannot surface for another, a hardened system prompt, and adversarial prompt-injection cases in the eval suite.

• Helped the teams I worked with use AI in their day-to-day delivery, improving the adoption measures tracked by their managers and helping them ship faster. Contributed to three of Pendo’s company-wide AI initiatives, including building AI features for the AI-native rebuild of Pendo Feedback. Named Pendo's AI MVP for Q2 2026.

The only project management platform built specifically for client work.

• Built Teamwork’s AI Profitability Forecaster, turning logged time, cost and billable rates into one-click revenue, cost and profit projections with confidence ranges. Self-hosted the models on AWS and routed each site by how much history it had: TinyTimeMixer above 512 days, Prophet between 90 and 511.

• Designed and built an AI onboarding flow that generated customised project plans, prioritised tasks and assigned work from user input, doubling onboarding completion rates.

• Built and shipped backend services in Go and PostgreSQL, taking ownership of the systems behind user-facing product features.

• Built user-facing features in Vue.js and ran growth experiments behind LaunchDarkly flags, using Pendo and HubSpot to measure results and roll out winners gradually.

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

• Joined at the zero-to-one stage, built Go and gRPC services on Google Kubernetes Engine, and integrated CardAlpha’s API with its partner bank’s API to create the company’s first automated merchant underwriting and onboarding flow.

• Rebuilt an existing iOS POS app as a web-based mobile app in React and TypeScript with Ionic and Capacitor, developing custom native plugins to connect phones to the bank’s payment terminals.

Empowering Futures with AI Innovation.

• Built backend microservices and REST APIs in Java with PostgreSQL and GraphQL, containerised with Docker, and automated build and deployment pipelines.

• Built large-scale web-scraping pipelines that collected gigabytes of data daily, then cleaned and prepared the data in Python and Pandas to train machine-learning models.

Seven tested, portable skills for coding agents, including senior-review for architecture-first review and agent-fix-loop for trace-verified reliability improvements.

Author of “Machine Learning and Its Use in Health Research”, a chapter covering machine learning, generative AI and Python applications in health research.

“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”: A post on the long-term talent-pipeline risks of AI hype in software engineering, viewed 2M+ times across LinkedIn and Reddit.

Graduated with a 3.11/4.0 GPA while working as a software engineer from my second year alongside coursework. Through the Ege Entrepreneurship Society, organised company visits and campus events that connected students with industry.