APPLICATIONS & DATA ENGINEER
Posted
Summary
Applications & Data Engineer with 8+ years of experience to build web apps and data pipelines using TypeScript/JavaScript and PostgreSQL. Key tasks include integrating REST APIs, designing data layers, building forecasting models, and implementing AI/LLM features using Supabase and Cloudflare. Requires experience with multi-site business data and fluent English.
APPLICATIONS & DATA ENGINEER
Hyderabad - WFO - Hybrid
Full Time
Australia Shift: 7-4 PM IST or 8-5 PM IST
8+ years experience.
KEY RESPONSIBILITIES
• Build and maintain internal web applications: reporting dashboards, and multi-user applications that store what people enter. That includes finishing the half-built ones, and moving working tools off personal machines and accounts onto client-managed hosting with the persistence, authentication and error handling they never had.
• Integrate REST APIs across point-of-sale, rostering, reservations, events, mobile ordering, accounting and workspace platforms, replacing manual exports with live feeds.
• Design and own the PostgreSQL data layer, including migration of historical data out of spreadsheets and platforms being replaced, and reconciliation across systems that disagree.
• Build scheduled pipelines that extract, reconcile, publish and alert, plus the daily and weekly reporting that runs on them. Run it in production: own the schedules, respond when something breaks, decide what happens when a source is late, missing or wrong, and follow up the discrepancies that do not resolve themselves.
• Build forecasting and inventory models: one framework configured per venue, validated out-of sample, with a manager reviewing and signing off the output. Managers often know something the model cannot, so an override is captured with its reason rather than treated as an error.
• Build Model Context Protocol (MCP) servers over vendor APIs where the platform ships no connector of its own, so an AI assistant can be pointed at a system like the point-of-sale directly, under the permissions of the person asking.
• Make the reporting interpret as well as present: what changed, why, and what to do about it. That means thresholds, analysis and, where it fits, a large language model whose output is checked before anyone relies on it.
HOW WE WORK
• One build per problem. Where several venues need the same thing, they get one application configured differently for each.
• Everything goes into client’s GitHub from the first commit and deploys through CI. Credentials never live in a repository. A project is finished when someone else can pick it up and change it, which makes documentation part of the work.
• The engineer is accountable for every line, including what AI wrote. What goes into production is what they have read, understood and can explain.
• Nothing goes live without a second set of eyes on the numbers: the department that asked for it checks the output, and the person who owns the number signs it off.
• Several of these systems hold customer and staff personal data. Access is scoped to who needs it, and nothing is exported somewhere it can no longer be controlled.
• IT is a small team in two countries. Overcommunicate early and ask before building the wrong thing. Take the question to the Head of IT, or go direct to the person who knows
Requirements
MUST HAVE:
Technical experience:
• Five or more years building web applications and the data pipelines behind them.
• JavaScript or TypeScript across both ends: browser and serverless.
• PostgreSQL, including schema design: taking spreadsheet-shaped data and giving it a structure that holds.
• Integrating commercial platforms over REST and OAuth 2.0, including at least one finance, ERP or point-of-sale system.
• Scheduled data pipelines in production that deliver their output by email and message, with error handling and alerting when a run fails.
• Role-based access control, and scoping data so each person sees only the venues they are responsible for.
• Serverless hosting and scheduled functions. Cloudflare and Supabase specifically, or close equivalents that transfer quickly.
• Git, with pull requests and deployment through CI.
• Working in an IDE with an AI coding assistant, under version control. Claude Code is the group standard.
• Integrating a language model into a product feature, and validating what it returns.
• Time-series forecasting or demand planning, validated out-of-sample. Funnel and cohort analysis. Comfort in financial data: revenue, cost of goods, labour, margin and reconciliation to the ledger.
• Supabase in production; Microsoft Entra ID, 365 and Graph; Azure SQL and Power BI, for helping retire whatever remains of the legacy environment.
• MCP servers, or similar tool interfaces built for a language model.
• Experience with multi-site businesses – venues, stores, clinics or branches that trade separately and report into one head office. Hospitality is ideal, but the pattern matters more than the sector, and experience in our particular platforms is not expected.
Beyond the technical:
• Reading code you did not write and judging whether it can be trusted – much of what arrives was built in a chat window by someone who is not a developer, with no meaningful version history, no comments, no tests, and numbers that have not always been verified.
• Fluent English, written and spoken – much of this role is explaining what has been built, or what is wrong with something inherited, to people who are not technical, including on calls with head office and venue staff.
• Working directly with the business, with no analyst in between – taking a brief in whatever form it arrives, and working alongside colleagues in another country and another discipline without being managed day to day.