Saveer Govindu

I take products from idea to the App Store, frontend to backend.

Full stack engineer in Melbourne. Two mobile apps live on the App Store, a platform migration and an AI support system shipped.

Currently at Les Mills International. Cofounder of Servy.

Portrait of Saveer Govindu

Work

Every project here runs from the screen down to the data, and I stay with it through launch.

Servy

A marketplace for home services. I cofounded it, built every layer and shipped both apps to the App Store, where they're live with real users.

  • Customer home
    Servy customer app home screen with search, active bookings and service categories
  • Choose a service
    Servy customer app list of available cleaning services
  • Booking flow
    Servy customer app booking flow, step one of three, choosing a date and time
  • Provider dashboard
    Servy provider app dashboard with weekly earnings, rating and next job
  • Provider earnings
    Servy provider app earnings screen with payouts and transaction history
Screen
Two React Native apps, customer and provider, updating in near real time.
API
One AppSync GraphQL API serving both the customer and provider apps.
Backend
92 Lambda functions, Stripe Connect payments and location search.
Sign in
Apple and Google sign in, so customers join without another password.
Testing
Used the Maestro MCP to turn user flows into reusable end to end tests across both apps.
Shipping
Used EAS pipelines to ship through Apple review and Google Play.

Servy NZ (customers) Servy Mates (providers) servy.co.nz

AI support system

Les Mills. Lead engineer.

AI that makes sense of Les Mills+ support tickets and comments. Ticket triage reaches 95% accuracy on live tickets, moderation false positives are down 90%, and 6,000+ tickets a month become insights for each team.

Triage
A fine tuned Vertex Gemini model trained on our tickets and validated with the support team, lifting accuracy from 70% to 95% on live tickets.
Reporting
A Lambda, SQS and Vertex AI pipeline that summarises tickets and generates customer voice insights, including the top 5 pain points for each team, built with Customer Ops.
Moderation
A fine tuned Gemini model for comment moderation that cut false positives by 90% compared with the previous model.

Content Management System

Les Mills. Lead engineer, coordinating a senior and a junior engineer.

A content management system for publishing workout videos, releases and everything else in the Les Mills+ app, including collections and programs of workouts. It replaced the legacy platform with zero customer downtime, and after launch I fixed 13 bugs with support and QA, cutting turnaround from weeks to days. Built as a Next.js app on Supabase, linked to AWS services and shipped behind feature flags.

Favourite Workouts

Les Mills. Key contributor.

A feature on the Les Mills+ website that lets members favourite workouts. I built it down the whole stack: the React website, the APIs behind it and the DynamoDB schema, and brought members' existing favourites across from legacy systems.

Also at Les Mills: React Native releases for LG, Samsung and Tizen TV and the mobile app.

About

I like owning a product from the first idea to the day real people use it, from the screen to the API to everything behind it. Open to product engineering roles in Melbourne.

AI first
I use Claude Code every day, with prompt engineering, agentic engineering and MCPs for Atlassian, Datadog, Maestro, Playwright, iOS Simulator and Context7.
Product first
I start from what the user needs at the end, then work backwards to the design and the code.
Production owner
I stay on after launch, working with support, QA and senior engineers on live issues, using CloudWatch and Datadog.

Got a product that needs to ship? Email me.