aeravel Ventures

For CTOs, heads of product and founders with a live product

Product engineering that starts before the ticket and ends after the release.

We work the way a forward deployed engineer does. We sit inside your team, shape a rough request into a pitch, put a prototype in front of people, build it behind a flag, then read the numbers to find out whether it landed. AI runs through all of it. The decisions that are expensive to get wrong stay with people.

Elixir · Phoenix · Rust · Go TypeScript · React · Angular Fintech · insurtech · commerce
Talk about your product See how we work
AI throughout plan pitch · sketch decide build prototype · flag measure events · data

The whole cycle

Most contractors take the middle third of the job

They arrive when the spec is written and leave when the branch is merged. That leaves you to decide what is worth building and to find out afterwards whether anyone used it. We take all six stages, because the first and the last decide whether the build was worth doing.

  1. Pitch

    We gather the problem with your product team and your customers, then set an appetite and the no-gos.

  2. Fat-marker sketch

    We agree the shape of the solution with your team before anyone opens Figma.

  3. Prototype

    A clickable demo, often built with AI in a day, that your team reviews before the real build starts.

  4. Grill the spec

    We stress-test the RFC in "grill me" sessions, tie each slice to a key result and list the events it must emit.

  5. Build behind a flag

    Each change ships behind a feature flag and rolls out tenant by tenant, with agent-run QA on staging.

  6. Measure

    Events feed dashboards that show whether the key result moved. If it did not, the point where users drop off becomes the next pitch.

How we run this with AI, and where a human still decides

Services

What we take on

Product engineering

Embedded in your team for the full cycle, from the first pitch to the dashboard that says whether it worked.

  • Pitches, specs and RFCs
  • Prototypes
  • Backend, APIs and web front ends
  • Payments and commerce integration
  • Instrumentation and analysis

AI in your delivery process

Putting AI into the way your team already works, at the steps where it pays, with review gates that catch its mistakes.

  • Repository style guides and conventions
  • Planning and scoping practice
  • Review gates and control points
  • Prototyping workflow
  • Team training on the above

Modernisation and review

Systems that still work but have stopped being safe to change, and the audits people ask for before they commit money.

  • Framework and runtime upgrades
  • Performance and infrastructure cost
  • Architecture review
  • Code and security audit
  • Technical due diligence

Each service in detail, with what it delivers

People

Who you work with

Open source

Public code instead of case studies

We have no client logos to show yet. The commit history below is ours, and it is public.

angularspree 1,714 stars · 810 forks An Angular storefront and admin for Spree stores. We are its top contributor by commits. gringotts 502 stars · 54 forks Several payment gateways behind one Elixir interface. We are its second contributor by commits. avia 536 stars · 109 forks An e-commerce framework in Elixir and Phoenix that uses Gringotts for payments.

All of it, with what we did in each

We also make small games: a daily word game for Elixir, React, AI, DevOps and travel, and a simulator of what skipping a stage of our process costs.

Start with the problem you have

A few sentences about the problem are enough. We reply with questions first.

Describe the problem