Sachin – React Native, Node.js, AWS

A performance nutrition platform for elite athletes and their coaches. It turns each athlete's training data from wearables and training apps into a personalised fuel plan for every meal, through an athlete mobile app and a web app for coaches and teams. It's used by 40% of the professional peloton at the Tour de France and 50% of the Premier League.

• Joined as a senior full-stack engineer and became CTO.
• Lead a team of four engineers and own architecture, infrastructure, security and budget, while still writing production code most days.
• Rebuilt the core PostgreSQL data model (years of history per user, millions of rows per table) with co-located JSON and new query patterns, cutting multi-day queries from several minutes to under 700ms, a key factor in closing a $2.1M seed round.
• Was, in the CEO's words, instrumental in building a product that delivered a 300%+ increase in ARR and found product-market fit.
• Led the migration from Supabase to a multi-region Google Cloud stack (EU, US and Australia) built with Pulumi, which also lowered infrastructure cost.
• Set up the SOC 1 and SOC 2 programs on Vanta and built the controls into infrastructure code: GCP-native alerting, VPC flow logs, bastion-only SSH and a public status page.
• Built CI/CD with multiple dev deploys a day, weekly production releases and same-hour hotfixes, with monitoring and rollback.
• Built backend services in TypeScript and Node.js (GraphQL, Prisma), a Python calculation engine and Kafka pipelines, integrating wearables such as Apple Health and Whoop.
• Shipped AI food logging with Gemini (photo, barcode and voice).
• Rolled out AI coding tools across the team (Claude Code, Codex, Cursor) with shared AGENTS.md rules, pre-commit checks and the same code review for AI-written changes.
• Hired and led the team through 1:1s, performance reviews, code reviews and incident response with blameless postmortems.

Freelance engagements through the Toptal network. The main ones: an agentic AI platform for financial due diligence, a HIPAA-compliant healthcare product, the mobile app and storefront search for one of the top 3 stores on Shopify, and an AI virtual try-on platform.

• Owned the architecture and shipped code myself across all projects, usually leading a small team.
• Built an agentic RAG platform for financial due diligence from scratch as CTO, using LangGraph, LangChain and LangSmith.
• Indexed 10M+ pages and scored 100% on the Vectara RAG benchmark, using hybrid retrieval (vector, BM25, HyDE) and an LLM verification layer behind page-level citations.
• Cut the cost per page of document processing by running every page through a cheap Gemini pass and only the pages with charts through a heavier one.
• Took the same startup through SOC 2 and owned its GCP infrastructure and hiring.
• Owned HIPAA-compliant authentication and authorization, the payments pipeline with reconciliation, and PHI-safe audit logging for a US healthcare company.
• Led the Fashion Nova React Native app (one of the top 3 Shopify stores, about 2 years), built on the Shopify Storefront GraphQL API.
• Led search, collections and catalog work on Shopify Plus, with hundreds of product launches a week.
• Built the mobile experience for joybird.com.
• Built microservices, public APIs and a developer dashboard for an AI virtual try-on platform.
• Built a React Native try-on app holding a steady 60 FPS with real-time garment detection on device.
• Cut deployment time from days to minutes with automated CI/CD.
• Kept test coverage above 70% with Jest and Appium.

A web-based 3D modeling tool for content creators making 3D content for brands and publishers like the NFL, HuffPost and Yahoo. Creators uploaded huge source files, and the platform turned them into optimized 3D assets that rendered well on every device, from desktop Chrome to iOS Safari and Android WebViews.

• Led a team of nine engineers (5 backend, 3 frontend, 1 platform) across time zones, with regular pairing and code reviews.
• Designed the backend from scratch as microservices on AWS: Lambda for stateless APIs, Kubernetes (EKS) for long-running processing, and event-driven queues between services instead of cascading synchronous calls.
• Built a server-side pipeline running Blender headless on Kubernetes, behind a queue with retries, that turned huge creator files into optimized 3D assets for every platform with no human in the loop, and cut the cost of each processed model.
• Developed a way to keep multiple databases in sync using streams.
• Wrote GraphQL microservices in TypeScript and Go with published docs, and kept synchronous responses under 300ms.
• Kept test coverage above 70% (Jest, Vitest, K6, Cypress) and built a Docker dev environment that cut onboarding from weeks to days.

A student-data dashboard for a network of public schools, used by staff at different access levels to manage student records and see what helps students succeed. It replaced an expensive legacy system.

• Volunteered as the lead engineer, recruited and led a small remote team of four engineers that shipped dozens of features on schedule.
• Built a dashboard for managing student data across user levels with React, Redux and Node.js, replacing a legacy system and saving more than $10,000 a year.
• Built data visualizations that helped staff find the factors that improve students' success.
• Built uploads that parse different data types into a single-source database, plus a backup and restore system and full data export.

A digital publishing platform where publishers turn PDFs into content for the web, mobile and social channels, read by more than 100 million unique users a month.

• Worked across the stack: creator tools and an HTML5 canvas PDF reader on the front end, and the OCaml services behind them.
• Developed a next-generation web PDF reader by breaking PDFs into their basic form, helping creators create and share content efficiently.
• Worked on OCaml backend entities serving north of 100 million unique users every month.
• Improved machine learning algorithms that extract articles and their content, reducing manual work for users.
• Created the technical approach for projects across all phases of the software development lifecycle.
• Worked with the development team to establish coding standards and monitored guidelines that give prompt feedback for timely adjustment and revalidation.
• Onboarded new developers to the codebase and got them productive.

District, a digital banking platform for business customers: one place to manage accounts, payments and liquidity across banks and currencies, with real-time balances, custom dashboards and a marketplace of third-party apps. Thousands of businesses used it daily, under GDPR and PSD2.

• Worked across the stack and wrote the SDK that other teams and outside developers built on.
• Developed a cross-browser financial platform for companies called District using reactive programming, Cycle.js and React.js.
• Wrote multiple libraries (SDK) for other teams to build for and communicate with the platform.
• Improved the developer onboarding experience on the platform and supported other teams within the organization.
• Made architectural decisions and kept a full-stack overview.
• Implemented API microservices and automated end-to-end tests.
• Built automated pipelines in GoCD that build, test and deploy code from commit to customers using new features.
• Responsible for deployments, releases, static servers, bugs and issues.
• Worked with centralized monitoring environments (Grafana, AppDynamics, Kibana) that gather system metrics.
• Interviewed, supported and mentored other developers.

Apps for Atlassian's Confluence and Jira, used by documentation and product teams to write, manage and publish content.

• Built React and Redux front ends for web apps, working closely with product owners and UI designers.
• Built back-end REST APIs on AWS Lambda and Node.js.
• Covered the whole cycle: prototyping, feature work, bug fixing, unit tests and automated functional tests.
• Gave tech talks and took part in hiring.