Construction bidding is a ~$2T U.S. industry, and the estimating layer that determines who wins projects is still largely run through email, spreadsheets, and manual vendor coordination.

I'm building Nilo — a system designed to turn vendor email traffic into structured estimating data.

Today the workflow between RFQs, vendor quotes, bid analysis, and execution feedback is mostly manual. Nilo focuses on structuring that process.

The MVP focuses on:

• RFQ management • vendor quote parsing from email/PDF replies • trade-level quote coverage tracking • structured quote comparison for estimators

The goal is to build a working MVP that can run inside real construction bids within ~6 weeks.

Looking for a founding engineer interested in building messy real-world systems involving:

• inbound email ingestion pipelines • document and PDF extraction • event-driven workflows • LLM-assisted parsing and classification • early-stage SaaS architecture

Equity: 1–5% (4-year vest / 1-year cliff). Equity-first during the MVP phase while the system is built and validated.

One challenge I'm thinking through is the quote-parsing pipeline.

Vendor quotes arrive in many formats: • email body totals • PDFs • Excel attachments • partial scope notes

The current design uses confidence scoring to decide whether to: • auto-create a quote • route to a review queue • treat the message as non-quote communication.

If anyone here has built systems that parse messy inbound email workflows or document extraction pipelines at scale, I'd love to hear what approaches worked well.