Agnese Picchio
Python Developer | LLM Pipelines, Document Extraction & Auto
New to Contra
Agnese is building their profile!
Built the ingestion layer for an AI SaaS platform as sole engineer. The pipeline takes messy multi-column PDF reports and outputs structured records: coordinate-based page cropping to drop headers and footers, whitespace normalisation to repair text fragmented across columns, and chunking sized for downstream processing. Extraction returns schema-validated objects, each claim paired with its verbatim source quote and page number, rather than free text, so the output is directly usable downstream. Bounded retry and fallback logic means a failed call degrades gracefully instead of corrupting the dataset.
0
3
Async pipeline ingesting a continuous BLE sensor stream through a queue buffer, applying a rolling-average filter and a PID controller to convert noisy input into smooth, stable output driving multiple hardware actuators concurrently, without blocking the control loop. The monitoring interface is decoupled via a JSON state file, so a UI fault can't stall the running session. Validated across 11 sessions and roughly 2,600 cycles, with a mean correlation of r = 0.976 between input and controller response.
0
7
Built an end-to-end pipeline against the UKHSA public health API and an interactive dashboard on top of it. Handles paginated fetching and nested JSON parsing, with a caching fallback so the dashboard keeps serving data through upstream outages (the reliability work turned out to be larger than the fetching). The interface uses async refresh and dynamic controls, updating without page reloads.
1
19