Retrieval systems in production, agent workflows that validate their own output, networks under real load and endpoint estates in the hundreds. Code walkthroughs are available on request during scoping.
A RAG chatbot that answers real IT-support questions from a company's own documentation instead of guessing.
Four business agents where the LLM does the reading and deterministic code does the checking: invoices, leads, content and support.
A retrieval system that classifies the question first, then picks between semantic search, a knowledge graph, or both.
A dual-agent system that attacks safety-aligned models on purpose, then measures exactly how and when they break.
A FERPA-aligned model that finds and removes personal data from student writing before it reaches a third-party system.
A generation system for a music-production client where every line is validated for syllable count, stress pattern and rhyme.
A proxy layer for online behavioural experiments that groups participants, blocks duplicates and routes everyone correctly.
An open-source library that explains *why* an imbalanced dataset is hard, not just how to resample it.
A flowchart canvas for designing system architecture that turns the finished diagram into scaffolding code.
Upload a CV, get the German university programmes you actually qualify for, with deadlines and contacts.
Managing and automating hundreds of Windows, macOS, iOS and Android endpoints across international offices.
Ticket routing, diagnostics and a knowledge portal that together removed the most repetitive half of a service desk's work.
A segmented network with firewall policy, remote-access VPN and intrusion detection: built, attacked and documented.
Identity and marketing site for a BPO and customer-support company, shipped on a modern stack.
A retrieval assistant over biology material that answers from the source text and shows where it came from.
Describe the problem in two lines. You get a plan, a fixed price and the engineer who would lead it, within 24 hours.
