Context

Over three years at BairesDev, first as a Sourcing Analyst and then as a Growth Automation Associate, most of my work followed the same shape: find a process that people were performing by hand, understand why it was manual, and replace it with something that runs on its own and can be audited afterwards. The specifics below are described at the level of the technical pattern rather than the internal implementation. The systems are proprietary; the approaches are not.

Partner registration pipeline

Problem. Partner registration data arrived as spreadsheets and had to be entered into an internal ERP by hand. Manual entry meant transcription errors, no record of what happened when, and no way to reconcile the two systems when they disagreed. Approach. A cloud ETL pipeline connecting spreadsheet intake to the ERP through its API, with a validation layer between them. The design priority was traceability: every record carries a log of what was ingested, what was transformed, and what was written, so a discrepancy can be traced to its origin rather than argued about. Outcome. Registration became continuous rather than batched, and the reconciliation problem largely disappeared because both systems were now fed from one validated source.

Survey automation

Problem. A recurring survey programme required someone to schedule it, launch it, collect responses, and trigger the follow-up campaign — every cycle, manually. Approach. An end-to-end workflow chaining form distribution, response collection, and campaign dispatch, so a monthly cycle runs start to finish without intervention. Outcome. The recurring manual effort went to zero and cycles stopped being missed, which is the failure mode that had actually motivated the work.

AI-assisted outreach

Problem. Identifying genuinely relevant engagement opportunities across a large volume of professional-network activity is a filtering problem that does not scale with headcount. Approach. Workflow automation with an AI evaluation step to surface high-value signals, feeding tailored outreach rather than volume-based sequences. Outcome. Outreach shifted from broad and generic to selective and specific, on a system that ran continuously instead of in bursts.

Earlier: sourcing operations

Before the growth automation work, as a Sourcing Analyst I automated recurring data entry with JavaScript and spreadsheet tooling, eliminating over 30 hours of manual work per week. I also built Python validation frameworks against internal APIs to improve data integrity, developed retrieval and search tooling, and administered the virtualised cloud environment the team's data operations ran on.