I take on 2-3 fractional CTO engagements at a time. I go deep with a few companies rather than shallow with many.
Who I work with
You might be a good fit if:
- You're a Series A or B company with traction and users
- Your engineering team is 5-20 people and growing
- You're hitting scaling challenges — hiring, architecture, delivery, AI adoption
- You need senior technical leadership but $400K+ for a full-time VP Eng or CTO doesn't make sense yet
Problems I Solve
"Our engineering team is a mess."
Delivery is unpredictable. People are frustrated. Nobody knows what "good" looks like anymore. I've walked into teams on fire and turned them around.
- Humi (Fintech): Stepped into the Director of Engineering team of 15 engineers for a payroll product moving $200M a month. The rate of bugs and support requests had been growing significantly and the team was drowning just doing support. I doubled the speed of delivery, payroll bugs dropped from 67% to 40% in 4 months. Implemented the leveling matrix for the entire engineering org which is still publicly available today.
- Triage (AI): Inherited a stalled engineering/ML org of 13 who had been working on an ML model for two years and no product. I stepped into the CTO role, identified that our model was great at doing a "hotdog / not hotdog" for Psoriasis, a very common skin condition. I oriented product, engineering and the exec around a product narrowly focused on that and launched the first ML-based dermatology app in 4 weeks. This was the precursor to the main product that was used over 500k times.
- Zapier: Inherited a data team with no strategic direction with 22 custom ingestion scripts that nobody trusted. Non-technical stakeholders couldn't access their own data without filing engineering tickets. I defined a real roadmap, replaced the scripts with Meltano (200+ open-source connectors maintained by the community), and made data self-service. The team went from firefighting to platform organization.
"We want to implement AI/LLMs."
Most companies are paralyzed by AI or shipping demos that don't work in production. I've shipped AI that actually works — reliably, at scale, without bankrupting you on inference costs.
- Spellbook (Legal AI): Their AI playbook agent had accumulated so much tech debt it was slow and expensive to run. I rebuilt the entire prompt architecture and token optimization — 50% faster, 60% cheaper, serving 8,000+ lawyers. In legal AI, hallucinations carry malpractice risk. I built an evaluation framework using lawyer-verified ground truth and automated regression testing. Accuracy went from 82% to 90%.
- Clio: The AI team had traditional ML backgrounds and was skeptical about GenAI's non-deterministic nature. No strategy, no investment, lots of hand-wringing. I built an OpenAI Word Add-in that helped lawyers draft and review documents. The POC was demoed to the CEO who endorsed the vision on the spot. That work became part of the foundation to Clio's AI Platform strategy.
"We need to ship faster and unblock sales."
Technical problems are blocking deals. Performance issues are losing you customers. I find the bottleneck and fix it.
- Spellbook: Enterprise law firms needed team-based document sharing, but the existing ACL permissions couldn't handle it. $3M+ in deals were sitting blocked in the pipeline. I led discovery on SpiceDB (Google Zanzibar-inspired authorization), built the POC, wrote the design doc that got cited internally as the "gold standard" for technical planning, and coordinated the implementation.
- Clio: Document sync was causing system lockups for mid-market law firms. Sales was losing deals to performance objections. I diagnosed the root cause, benchmarked three solutions in a single day, and led the implementation. 5-10x speed improvement. The objection disappeared from the sales cycle.
- Humi: Payroll tech debt was slowing everything down, but leadership wouldn't invest without business justification. I quantified the debt in dollars — $304K/year in lost engineering capacity while simultaneously delivering new hourly payroll features the business needed. Scaled onboarding capacity from 1,600 to 3,500 companies.
"We want to build a new product line."
You have a core business that's working. You want to diversify. I've done this multiple times.
- Aerial.ai: WiFi sensing technology that could detect presence and movement without cameras — privacy-preserving, genuinely novel. But the CTO had a technical demo, not a product. I turned it into product specs and a GTM strategy, then led direct sales with major technology companies. Closed 8 strategic partnerships worth $150K+ with Amazon, Samsung, Sony, Panasonic, and Videotron.
- Clio: Product wanted GenAI in Microsoft Word but faced uncertainty about feasibility and UX. Weeks of async discussion, nobody making decisions. I organized an intensive two-day working session and led 3 rapid POC iterations, each de-risking a different aspect. The PM said: "These two days did what we otherwise would have taken weeks." Multiple blocked initiatives got unblocked simultaneously.
How I Work
Relationships first. I'm not the consultant who hides behind frameworks and slides. I build real relationships with the people I work with.
Start with listening. Before I tell you what to do, I need to understand what's actually happening, not just what leadership thinks is happening, but what the engineers experience.
Have the hard conversations. A lot of engineering problems are actually people problems that nobody wants to address. I come from the school of Crucial Conversations as my framework for finding common ground to move forward together.
Make it stick. I'm not trying to create dependency. I work with you to put things in place for lasting change: finding the right people for your team, processes, career frameworks, technical patterns. These are the things that keep bringing you to the next level after I'm gone.
Track Record
20 years in software. 8 years managing teams of 8-14 engineers.
Spellbook · Clio · Humi · Zapier · Aerial.ai
I've also done consulting work for organizations navigating AI adoption in high-stakes environments where getting it wrong has real consequences.
What I Don't Do
- Pre-traction companies. If you have an idea but no users, no revenue, and no proof that anyone wants what you're building, that's not me.
- Substitute for a technical cofounder. If you're a non-technical founder at the idea stage, you need a cofounder or an agency.
- Pure strategy. I don't write 50-page roadmaps and hand them off. I stay involved in execution.
- Manage purely offshore dev teams. I work with companies building real internal engineering teams.
Let's Talk
If that sounds like where you are, email me: jevin@quickjack.ca