Michał Kijowski

Part of the team building Astrix's Non-Human Identity security platform — discovering and securing exposed credentials and secrets across an organization's systems, and more recently extending into AI agent and MCP security.

• Developing and maintaining a secret-scanning solution for Non-Human Identity (NHI) management
• Building AI agent security detection — surfacing AI agent usage patterns and vulnerabilities from EDR telemetry
• Building scalable backend services to detect and analyze secrets across multiple data sources
• Improving the team's AI-assisted development workflow — authoring Claude Code hooks, skills, and MCP servers used in day-to-day engineering
• Stack: Python, TypeScript, PostgreSQL, MongoDB, AWS

Worked on the backend systems responsible for scraping and analyzing large volumes of analytical data.

• Extended a GraphQL-based backend for analytical data
• Implemented several web scrapers for new complex data sources
• Worked closely with data engineers to implement their models in production systems
• Stack: TypeScript, Node.js, GraphQL, PostgreSQL, Google BigQuery

Took a few months off to spend time with family and relocate between two cities.

At an early-stage startup, I focused on developing and improving the company's core algorithms.

• Greatly improved core algorithm performance, reducing calculation time tenfold.
• Invented a unique statistical model to assess user health based on the analysis of their in-app actions.
• Stack: TypeScript, Node.js, TensorFlow, MySQL, Google BigQuery

Led the ongoing maintenance and enhancement of a remote moderated UX testing tool, originally built at Validately.

• Led the internationalization of our moderated application.
• Hired and onboarded a strong developer who quickly became a productive member of the team.
• Stack: TypeScript, Node.js, Redis, MySQL, WebRTC, WebSockets, React, Redux, GCP, AWS

As lead developer of Validately's moderated UX solution, I planned and led the integration of the two systems following the merger with UserZoom.

• Co-led the integration of Validately's moderated UX solution into UserZoom's platform following the acquisition, making it available across both UserZoom and UserZoom GO (formerly Validately).
• Stack: TypeScript, Node.js, Redis, MySQL, WebRTC, WebSockets, React, Redux, GCP, AWS

As a full-stack developer at an early-stage startup, I took responsibility for many parts of the system, focusing primarily on core customer-facing applications.

• As lead architect and implementer, built a remote moderated UX testing tool — a React frontend, Express/Socket.io backend, and a Google Chrome extension, all in TypeScript. Praised by management and clients for its stability and robustness, it began as a solo effort, grew into a team-maintained product, and became a key factor in the company's acquisition by UserZoom.
• Designed and developed an unmoderated UX testing tool, a Google Chrome extension with screen recording features.
• Stack: TypeScript, Node.js, Redis, MySQL, WebRTC, WebSockets, React, Redux, GCP, ffmpeg

In my role as a Software Developer within the Quantitative Analysis Department at Orbis Investment Advisory, I contributed significantly to several important initiatives.

• Played a pivotal role in the design and implementation of an advanced analysis tool for quantitative analysts, which leveraged Reactive (Rx) and Dataflow Programming. This tool, essential for backtesting and evaluating alpha models during portfolio creation, became the favored choice for several analysts.
• As lead implementer in a three-person team, engineered a tool to help portfolio managers optimally manage their portfolios' cash flows.
• Stack: C#, SQL Server, WPF, WCF, Matlab, R

Played an active role in the Synat research initiative, a project funded by the National Research and Development Center's strategic programme.

• Developed an innovative semantic faceted search engine based on data extracted from Wikipedia, as my master's thesis.
• Co-developed an efficient relational database schema that streamlined storing and querying parsed scientific articles.
• Stack: C#, Python, PostgreSQL

GPA: 4.5/5

Thesis: Implementation of a system for conceptual graph queries

Main focus: knowledge representation, natural language processing, information retrieval, distributed and parallel programming, compiler construction

GPA: 4.5/5

Thesis: Using faceted search for navigating through scientific documents

Main focus: data mining, machine learning, natural language processing, information retrieval

GPA: 4.5/5

Thesis: Co-Training approach in Machine Learning

Main focus: data mining, machine learning, statistics