Runs the nine diagnostic checkpoints across the seven-layer growth system, ranks the constraints supported by evidence, and writes the costed plan every engagement starts from.
About Ariadne
A growth engineering firm, and why it exists
Ariadne is a growth engineering firm for owner-led service businesses. Its teams design, build, and instrument the connected systems behind revenue: lead generation, conversion, CRM, follow-up, automation, and measurement. The company was founded by Sourena Khanzadeh and is organised around capability teams rather than a menu of unrelated retainers.
Toronto, Ontario, Canada · Remote delivery across Canada and the United States
Toronto, CanadaSoftware engineer · AI researcher
Origin
Why growth engineering rather than an agency.
The observation that started this: many business problems that look like marketing problems are actually systems problems, and systems problems are an engineering discipline.
A business buys a website, then an SEO retainer, then a CRM, then an automation tool. Four vendors may be measured on their own pieces while the end-to-end customer outcome has no clear owner. The gaps between those pieces are where money can be lost: the website-to-CRM handoff may have no clear owner, a stopped follow-up sequence may go unnoticed, and an SEO retainer may produce traffic without revenue attribution ever being wired in.
Ariadne exists to build the pieces and the gaps, in that order. The method is the same one used on any complex system: understand how it currently behaves, find the constraint, change one thing, measure what happened.
The name comes from the thread that made a labyrinth navigable. That is roughly the job: map the path, then make it something a business can actually follow.
Structure
Capability teams, not channel silos.
Ariadne is organised around the layers of a growth system rather than around channels to sell. An engagement draws on whichever teams the diagnosis calls for, under a single owner of the end-to-end outcome.
Builds the sites and conversion paths, wired into the CRM and analytics rather than shipped as a standalone brochure. Source code and accounts stay with the client.
Configures pipelines around how a business actually sells: real stages, entry and exit criteria, routing, and follow-up that keeps working when the week gets busy.
Fixes the process first, then automates it — with visibility, failure alerts, and a practical manual path for when something upstream breaks.
Assesses where a model is genuinely worth using and where it is not, and builds the ones that survive contact with real data. The same group publishes our open research.
Search, answer-engine visibility, and lead generation systems, sequenced after the downstream path can hold what they deliver.
Instruments the path end to end so the revenue question has an answer that is not a guess, using the minimum data needed to answer it.
Monitors the systems after launch, catches the silent failures — the stopped workflow, the broken integration, the lead that arrives at 2am — and keeps them maintained.
Holds the engagement together: scope, sequencing, handover documentation, and a named engineer who stays accountable for the outcome rather than a channel.
Founder
Sourena Khanzadeh.
Ariadne was founded by Sourena Khanzadeh, a software engineer and AI researcher based in Toronto. He builds full-stack applications, automation systems, and applied AI tools, and his public work spans software engineering, trustworthy AI, cybersecurity, and smart-contract optimization. He sets the engineering standard the delivery teams work to.
That background matters here for a specific reason. Growth systems fail in the same ways software systems fail: silent errors, unhandled edge cases, integrations that break without telling anyone, and no instrumentation to reveal any of it. The lead that arrives outside business hours, the duplicate submission, the workflow that stopped firing without an alert. Engineering has well-understood answers for that category of problem, and many conventional marketing engagements do not apply them consistently.
The background claims below link to public work or professional profiles that you can verify directly.
Verifiable
Public work and code.
Open-source Python workflow that tests whether an AI agent’s stated reasoning actually drives its answer. Captures reasoning, applies controlled interventions, reruns, scores, and exports structured results.
PaperAriadneXAI preprintThe research paper behind AriadneXAI, available as a preprint on arXiv.
PortfolioFounder portfolioSourena Khanzadeh’s software engineering, applied AI, cybersecurity, and smart-contract work.
ProfileGitHub profilePublic repositories and contribution history.
ProfileLinkedInProfessional history and background.
Working with us
How we work.
You work directly with the engineers doing the diagnosis and the build. There is no relay through an account manager who was not in the room when the decisions were made, and no handoff to a junior team after the pitch.
We staff engagements deliberately rather than filling capacity, so we will tell you our availability before you commit rather than after. Where a project needs specialist capacity outside our teams — paid media, video, high-volume content — we either bring in partners we have worked with and manage them, or we tell you to hire them directly and stay out of the markup.
What we say no to:
- Pre-revenue businesses with no existing demand to route.
- Projects where the goal is a cheap website and nothing after the form submit.
- Ecommerce and DTC. Different mechanics, different stack, and a specialist firm would do better work.
- Regulated projects that expect Ariadne to act as legal or compliance counsel, replace an EMR or practice system, or handle protected or privileged data without the appropriate agreements and client-reviewed constraints.
- Guaranteed rankings, guaranteed leads, or guaranteed revenue. Those claims are not defensible.
You own everything at the end: source code, accounts, content, domain, data, and documentation, with Git history intact so another capable engineer can pick it up.
Contact
Get in touch.
The most useful first step is a growth system audit. If you would rather just ask something, email support@ariadne.fyi and it reaches the team directly.
You can also read the full explanation of growth engineering, see the public artifacts, or start with the nine-checkpoint diagnostic and run it yourself.
Work together
Start with the diagnosis.
A growth system audit reviews nine diagnostic checkpoints across the seven-layer system, ranks the two or three highest-priority constraints supported by the available evidence, and gives you a costed plan you could execute with anyone.