[Hiring] Applied AI Engineer @| HCL Global System

Role Description

Forward Deployed AI Engineer, Intelligent Document Processing and Enterprise Data Matching

• Co-design and build document-ingestion + entity-matching + enrichment pipelines against a named owner and workflow.
• Integrate with enterprise authorities via API, event streams, and extract tables; design human-in the-loop review and exception handling.
• Instrument match quality (match rate, precision, manual-touch reduction, error rate) so results are measurable.
• Commit all artifacts to EBSCO repos as you go and hand off against a defined independent operation bar.

Qualifications

• Intelligent Document Processing - document/structured extraction (OCR and LLM-based), handling messy real-world inputs (Excel, PDF).
• Entity resolution / fuzzy matching / record linkage against reference data; data-quality and reconciliation experience.
• Strong Python and SQL; data pipelines; REST/API and event-stream integration; cloud (AWS preferred).
• Human-in-the-loop workflow design; ability to embed with finance/ops owners and transfer the capability.

Requirements

• Strong Eastern/Central Time overlap is required for daily embedded co-build (EBSCO hubs - Ipswich, MA & Birmingham, AL).
• Engagement shape: Time-boxed engagements (typ. 4-6 weeks build + hand-off overlap), 2-3 projects running concurrently.
• Seniority: Mid-to-senior applied engineers who can build production-quality work and teach it to an owner in parallel.
• AI-Driven Invoice Management (AP/AR) - invoice intake and extraction, then validation/matching of line items against EBSCO's Invoice/Product Authority, with human-in-the-loop confirmation.
• Federal Bids / Title Authority matching - read an Excel bid/renewal input, fuzzy-match each line against Product Authority (PA within PAVA on the ESOS platform), and auto-populate missing fields (ISSN, product keys, price, offerings).
• PAVA price-request automation - the inbound return (publisher price sheet → match to authority → enrich rates), today manual.
• Build the underlying reusable "match-and-enrich against Product Authority" service that all three share.

Benefits