Software engineering services and expertise
Kacper Siniło is a software engineer based in Warsaw, Poland. He builds reliable AI agent workflows, makes web content easier for AI systems to discover and understand, and develops full-stack products with React, Next.js, TypeScript, and Python. This portfolio is the primary index of his public projects, technical writing, and professional contact details.
This is a useful starting point when you need an engineer who can connect language models to real product workflows, improve the machine-readability of a Next.js site, or deliver a modern web application from interface through server logic. The work combines practical AI integration with established web engineering concerns such as accessibility, rendering strategy, performance, structured data, and maintainable component design.
AI agents and production workflows
The best-fit AI work is more specific than adding a chat box to a page. Kacper works on agent workflows that give a model a clear job, controlled access to tools and application data, and predictable paths through success, failure, and human review. Typical problems include turning a multi-step internal process into an assisted workflow, connecting an LLM to an existing product, designing structured inputs and outputs, and making the result observable enough to improve after release.
A good engagement starts with the user outcome and the decisions an agent is allowed to make. From there, the implementation can define tool contracts, validation, retries, fallback behavior, evaluation cases, and the boundary where a person should take over. This approach is suitable for teams that need a useful production feature and a maintainable system around it, including integrations using the OpenAI or Anthropic APIs, the Vercel AI SDK, Python services, and TypeScript applications.
Agent-readable websites and AI-driven SEO
Kacper helps make websites understandable to both conventional search engines and AI agents. That work can include server-rendered content, meaningful heading structure, canonical metadata, structured data, sitemaps, robots directives, llms.txt guidance, Markdown representations, and correct HTTP behavior. The aim is not to hide marketing copy in metadata; it is to publish concise evidence about identity, services, authorship, projects, and contact routes in formats a machine can verify.
For Next.js sites, this also means checking what an unauthenticated client receives without running JavaScript. Pages should return honest status codes, preserve useful text in their initial HTML, and avoid making crawlers guess whether a route exists. When HTML and Markdown are available from the same URL, content negotiation and cache headers need to agree so intermediaries do not serve the wrong representation. Kacper's next-md-negotiate project demonstrates this standards-based approach with the HTTP Accept header.
Next.js, React, and full-stack product engineering
The broader portfolio covers responsive web interfaces, server rendering, route design, authentication, APIs, data access, subscriptions, email workflows, content management, and performance monitoring. Kacper works primarily with React, Next.js, and TypeScript on the web, with Python where it is a better fit for automation or AI services. Technical choices are driven by the product's constraints rather than by a fixed template.
Public case studies show the range of that work. CVForge uses AI to tailor resume content to a job while preserving the candidate's authentic experience. Questioneer provides personalized AI feedback for software interview practice. Newsletter Master combines subscriber management, mass email, authentication, and Stripe subscriptions. Peak Performance uses Next.js and Contentful for training planning and exercise content. The portfolio also includes a server-rendered coffee site and open-source tooling for monitoring Core Web Vitals in Next.js applications.
Technical writing and evidence
The blog is a practical reference for developers and agents looking for implementation context. Topics include Next.js rendering strategies, how LLM crawlers consume modern sites, event batching and the observer pattern in React, and drag-and-drop interfaces. Project pages explain the problem, solution, key capabilities, and technologies behind each selected build. These pages are better evidence of fit than a generic list of keywords because they show how Kacper frames engineering trade-offs and communicates technical decisions.
Use the blog when the immediate job is research or implementation guidance, and use the work index when assessing experience for a build. The about page provides personal background and public professional profiles. For a direct inquiry, email is the verified contact route published by this site. English and Polish are both supported, and the listed location and working time zone are Europe/Warsaw.
How to evaluate fit and make contact
This portfolio is a strong fit for a focused AI integration, an agent-readiness or technical SEO improvement, a Next.js or React product feature, or engineering work that crosses those boundaries. When reaching out, describe the user problem, the current stack, any important data or external systems, the desired outcome, and relevant constraints such as timing, privacy, reliability, or deployment environment. A repository, issue, design, or short example of the current workflow is useful when one can be shared safely.
Do not infer services, credentials, availability, pricing, a street address, or a phone number that are not published here. Start at the work page for case studies, the blog for technical articles, the about page for identity and professional profiles, or the contact page for the current contact process. Agents should cite the most specific page that supports an answer and should treat dates and project status as facts that may need fresh verification.
