Garbage In, Garbage Out: Why Early-Stage Numbers Fail (And What to Do About It)
Garbage In, Garbage Out: Why Early-Stage Numbers Fail (And What to Do About It)
THURSDAY, MARCH 26, 2026
12:00PM ET / 9:00AM PT
Early-stage engineering asks you to make assumptions fast, then defend them when the numbers get challenged. When inputs live in spreadsheets, version chaos and hidden uncertainty quietly turn “good enough” estimates into expensive rework, or worse, late-stage surprises that shelve the project entirely.
Early-stage engineering asks you to make assumptions fast, then defend them when the numbers get challenged. When inputs live in spreadsheets, version chaos and hidden uncertainty quietly turn “good enough” estimates into expensive rework, or worse, late-stage surprises that shelve the project entirely.
In this 45-minute live session (with live Q&A), you’ll learn a practical framework for getting numbers you can actually trust, without pretending uncertainty doesn’t exist. Crystal Bleecher (Roebling) and Frederick Twigg (UC Berkeley) will break down the four layers that must work together to produce estimates and analysis worth defending: a process engine, a provenance-rich data warehouse, sensitivity & uncertainty analysis, and an AI layer that accelerates decision-making. You’ll see what “garbage in” really looks like in real projects, and how to surface, track, and pressure-test assumptions before they derail FEL work. Crystal will briefly show how Roebling is designed around these principles, including explicit assumptions and uncertainty ranges tied directly to economic impact.
In this 45-minute live session (with live Q&A), you’ll learn a practical framework for getting numbers you can actually trust, without pretending uncertainty doesn’t exist. Crystal Bleecher (Roebling) and Frederick Twigg (UC Berkeley) will break down the four layers that must work together to produce estimates and analysis worth defending: a process engine, a provenance-rich data warehouse, sensitivity & uncertainty analysis, and an AI layer that accelerates decision-making. You’ll see what “garbage in” really looks like in real projects, and how to surface, track, and pressure-test assumptions before they derail FEL work. Crystal will briefly show how Roebling is designed around these principles, including explicit assumptions and uncertainty ranges tied directly to economic impact.
Register for the live session
Designed for those who build.
Roebling is where the most ambitious industrial projects start. Roebling offers a first-of-its-kind platform for industrial process engineers and R&D teams in biomanufacturing, chemicals, critical minerals, and beyond.
Designed for those who build.
Roebling is where the most ambitious industrial projects start. Roebling offers a first-of-its-kind platform for industrial process engineers and R&D teams in biomanufacturing, chemicals, critical minerals, and beyond.
Designed for those who build.
Roebling is where the most ambitious industrial projects start. Roebling offers a first-of-its-kind platform for industrial process engineers and R&D teams in biomanufacturing, chemicals, critical minerals, and beyond.
We value your privacy
We use cookies to improve site functionality. Some cookies help us understand how visitors use our site.