Salinash Studio is the product engineering practice of Ali Siddique — designing, building and operating production software end to end, from ML inference infrastructure and GPU rendering to iOS apps and real-time web interfaces.
One exit via Acquire.com. Six iOS apps shipped. £50K+ in product revenue. AWS Certified Solutions Architect.
Three years of independent practice, measured by what shipped and what it did.
8 case studies — ML infrastructure, iOS and graphics, cloud platforms and products taken to exit.
An end-to-end pipeline that turns text and video into rig-ready animations for Roblox characters, served by motion-generation models running on serverless GPUs.
A native iOS viewer that parses DICOM medical imaging files and renders scan data through a custom GPU pipeline built on Metal.
A browser-based 3D tool that parses and converts models between OBJ, FBX, glTF and STL, and renders them interactively in real time.
Six SwiftUI applications architected, launched and grown solo — 20,000+ downloads, £50K+ in revenue and a 35% retention improvement.
Four disciplines, each proven on shipped work rather than listed on a skills page.
Serving models in production — separating ML compute from the web tier so GPU capacity scales with demand instead of sitting idle.
Native apps with custom rendering underneath — six shipped, £50K+ in revenue, 20,000+ downloads.
TypeScript front ends that stay responsive under load, including 3D and live-processing interfaces.
The infrastructure under the product — provisioned, observable and cheap to deploy to.
Start from the architecture, not the screens. Map the data, the compute, the third parties and the failure modes before a line of product code gets written.
Ship one path end to end — request intake through to rendered output — so the hard parts are proven early rather than discovered late.
Tests with Jest and Playwright, monitoring with Prometheus and Grafana, and a CI/CD pipeline that makes deploying boring.
Get it in front of users, then move the numbers — the retention, acquisition and content work that took past products to 4,000 users and an exit.
An AI-powered landscape design product taken from first commit to 4,000 users and sold through Acquire.com — full lifecycle, including the handover.
A practice platform for the game-based hiring assessments used by employers — HireVue and pymetrics style — grown to 500 users.
A real-time transcription interface with embedding-based semantic search across transcripts and AI-generated quizzes, tuned to stay smooth during live processing.
A Go microservice platform on Kubernetes with Redis caching, RabbitMQ queues, Prometheus/Grafana monitoring and a CI/CD pipeline that cut deploy time by 70%.
Reach the studio at [email protected]