Founding Engineering Lead

These are the real problems, not bullet-point abstractions.

Designing agent systems that ingest unstructured tender documents (PDFs, scanned drawings, BOQs, technical specs) and produce structured cost estimates that procurement teams actually trust enough to bid on.

Building a vendor-quote parsing pipeline that handles every format vendors send, PDFs, scanned images, Excel files, quotes pasted into email bodies, WhatsApp screenshots, and normalises them into comparable line items.

Architecting a multi-tenant platform where one EPC''s project data, vendor relationships, and pricing intelligence never leak across organisations, while still enabling cross-deployment AI learning in a privacy-safe way (this is our long-term moat).

Building the connector layer that lets Workwise ingest data from email, Google Drive, OneDrive, WhatsApp, TReDS, and major ERPs (SAP, Oracle, Dynamics, Tally, Strategy ERP, etc.). The product needs to work despite people, not because of them.

Deciding when to use deterministic logic vs. LLM reasoning vs. human-in-the-loop, and building the eval infrastructure to know whether you got it right.

Designing for modular deployment, Company A may want only Tender + Estimation. Company B wants the full lifecycle. Company C wants on-prem. The architecture has to support all of this without forking.

Hardening the platform for enterprise audits, encryption, access control, audit logs, on-prem deployability, compliance posture.

Setting the engineering standards for the team, code quality, system design, reliability, AI-assisted development workflows, how we ship.

A fuller view of the platform vision is in our internal Workwise V2 system direction doc, which you''ll be taken through in the process.