Unicornis Consulting LLP
I design, build, and review software, with a focus on AI agents, developer tools, and the systems they connect to.
My work brings together practical AI engineering and security research: making tools useful, defining what they can access, and understanding what happens when things go wrong.
An independent engineering practice · Working remotely with teams worldwide
Agents, MCP integrations, APIs, and workflows that connect models to useful work. From an early implementation to evaluation, inference infrastructure, and integration with an existing product.
Shared context, user permissions, secret handling, and tool execution. Working through who can see information, which actions an agent can take, and where human approval belongs.
Security reviews, threat modelling, and hands-on engineering across applications, infrastructure, and blockchain systems. Clear findings and practical guidance for the team doing the work.
I start with the system, the people using it, and what a costly failure would look like. Engagements can be a focused review, a specific build, or an ongoing engineering partnership. We agree on scope, responsibilities, and what a useful result looks like before work begins.
AI engineering for software security
In my consulting work with FYEO, I contribute to its AI security scanner across MCP and API integration, scanner prompts, model fine-tuning, and supporting engineering.
The work connects security analysis with the interfaces and workflows developers and AI assistants use to request scans and work with the results.
An AI agent with shared context and access management
I’m building Bender to support collaborative engineering workflows. It brings together shared context, multi-user access management, isolated secrets, and containerized tool execution, with support for local and hosted models.
The central design problem is making shared context useful while respecting each user’s permissions and controlling the actions an agent can take.
PIQ →Creator · In development
Software for climbing and training: finding board problems, logging climbs, and connecting training sensors and board lights.
A travel publication I work on.
Unicornis Consulting LLP is an independent engineering practice led by .
The practice draws on more than 15 years of software engineering, from building and operating business infrastructure to security research and AI systems, across backend services, developer tooling, distributed systems, and blockchain security.
Registered in British Columbia, Canada · Working remotely with teams worldwide
Contact · Unicornis Consulting LLP
Tell me a little about the project, the problem you’re working through, and the kind of involvement you have in mind. You don’t need a finished specification to start a conversation.
A plain-text overview covers the practice, selected projects, and how to get in touch. Share it with your AI assistant before starting a conversation.
https://unicornisstudio.com/llms.txt