Senior engineering, on the inside of your team.
Kernwise is a small consulting practice for the engagements other firms route around — architecture that will not hold, platforms that have outgrown their first design, and AI systems that need to ship without breaking the product underneath them.
- Engagements
- Architecture review, platform engineering, backend systems, AI integration.
- How we work
- Embedded alongside your team. Usually two to six weeks.
- Where
- United States · United Kingdom · Australia. Remote-first.
- Founded
- 2026.
Five engagements we keep being asked to take.
We do not do strategy decks or staff augmentation. We work on the parts of the system that are load-bearing.
Architecture review
We read your system for two weeks and write down what it actually does, then tell you where the next failure is going to come from.
Backend systems
We trace the query paths your backend actually runs, find where the data model stopped matching how it is used, and propose the smallest change that holds at the next order of magnitude.
Platform engineering
We measure what it actually costs to get one change into production, then remove the slowest and scariest steps and leave your team able to run it without us.
AI integration
We put a model into your product the honest way — by deciding what good means, measuring it, and shipping the smallest model that clears the bar.
Fix-it engagements
Something specific is broken in production and the team needs it found and fixed, not described. We reproduce it, find the cause, fix it, and leave you something you can maintain.
If your engagement is not on this list, it is probably one of these in disguise. Write to us and we will tell you straight.
Three things we agree before we start.
Not a process diagram. The actual structure of every engagement we take.
One person who owns the work, end to end.
No account manager between us and the engineering team. The person you talk to in the first call is the person writing code in week three.
002A written brief in week one.
By the end of the first week we hand back a brief in your voice — the problem as we understand it, the constraints, and the calls we would make if we were running it.
003A handoff document, not a handover meeting.
When the engagement ends we leave behind something the next on-call engineer can read at 2 a.m. and understand. Architecture, decisions, runbook, the things we got wrong.
A few of the rooms we have worked in.
Client names withheld. The outcomes are not.
Cutting p95 checkout latency to 340 ms
A checkout that took 1.4 seconds at p95 was losing orders at the last step. Two slow spans accounted for 93 percent of it. Neither was the database.
Cutting unsupported answers from 9% to 0.4%
A support assistant was answering confidently and wrong, inventing policy details that were never in the docs. The model was fine. The chunker was splitting policy tables mid-row, so the row the model needed never reached it.
From 41 minutes to 6, and nobody is afraid anymore
A deploy took 41 minutes and everyone dreaded it, so the team batched a week of changes and shipped on Fridays. Three slow steps were doing the damage. None of them needed a new tool.
Have a system that is holding, but only just?
Write to us at hello@kernwise.com. A real person on our side reads every message — usually within a business day.
If you would rather start from a form, the contact page is one screen and asks five questions.