A career in questions
One throughline. Many chapters.
Explore the work from imaging to AI. The connections are present-day reflections; each chapter carries its sources.
01See1993 record
Siemens Medical Systems · Diagnostic imaging
Learning to see.
In 1993, an AIA/LA program listed me as a Siemens Medical Systems speaker on future radiology equipment and planning. Looking back, I see a starting point for a recurring question: what changes when a system extends what a person can see?
Behind this chapter
The archival program supports the Siemens affiliation and speaking topic. The connection to later work is a present-day interpretation. Image Enhancement Systems and the broader early imaging chronology remain open.
Open source record ↗02Build1990s–2000s & teaching
Neoforma · Gabriel Venture Partners · Berkeley
An idea has to meet the world.
Neoforma went public in 2000. My work as an entrepreneur-in-residence at Gabriel Venture Partners and as a Berkeley lecturer put the same problem in different rooms: what do you do when a compelling idea meets an uncooperative reality?
Behind this chapter
The accepted Berkeley SCET source documents Neoforma, the Gabriel Venture Partners EIR role, and lecturer context. The chapter’s question is a present-day reflection, not a quotation.
Open source record ↗03Preserve2008 coverage · 2010 application
iFOREM · Digital preservation
What should outlast the system?
At iFOREM, preservation meant asking what happens to digital assets when the company holding them disappears. The architecture put an independent trust beside the service provider. I return to that separation today: the service and what must survive it are different responsibilities.
Behind this chapter
The accepted 2008 iFOREM coverage supports the preservation architecture and Gabriel backing. US20100138311A1 supplies the software escrow artifact. The final sentence is present-day reflection.
Open source record ↗04Structure2012–2013
GuessN · Human feedback
Give a question a structure.
At GuessN, the work included structured propositions, target audiences, and online community feedback. Today, I read that artifact through a question I still care about: how do you make a person’s meaning usable by a machine without quietly replacing it?
Behind this chapter
US20130232426A1 lists Stephen Pieraldi among four inventors and GuessN LLC as original assignee. Published September 5, 2013. The modern connection is retrospective, not evidence that this work anticipated generative AI.
Open source record ↗05Heal2013–2014
CuraSeal · Medical devices & patent families
Invention with human stakes.
My inventor record at CuraSeal includes fistula and wound-treatment applications. Here the design problem meets tissue and biology. For me, it is a useful discipline to carry back into software: a convincing description is only the beginning of a design.
Behind this chapter
US20140379025A1 and US20140379026A1 are published applications. The former lists Stephen Pieraldi among three inventors, Curaseal Inc as original assignee, and an international family including WO2014205269A2. No clinical efficacy or issued U.S. patent claim is made.
Open source record ↗06Protect2014 onward
Sideband Networks · Fidelis Cybersecurity · E8 Security · HP
Trust has to be designed.
From Sideband and Fidelis to E8 and HP, the story moves through security, machine learning, and the boundaries of trust. In 2017, the public record places me in AI and machine-learning strategy at E8. The current inquiry spans chip-to-cloud security, Zero Trust, workforce experience, and enterprise AI.
Behind this chapter
The accepted Sideband release supports VP Engineering; the accepted 2017 CSO source identifies the E8 role. Fidelis and the HP themes are supplied career context, with detailed artifacts still open. The HP–OpenAI partnership is organizational context, not evidence of personal leadership.
Open source record ↗07Reason2021–2026
Medium · LinkedIn · Pieraldi.com · YouTube
Think in public. Keep revising.
Writing is where I keep the argument open. From “Aspirational vs. Artificial Intelligence” to “State of AI and I,” the inquiry moves from AI’s promises and limits toward the experience of working alongside it. The essays below are a selective trail through that work.
Behind this chapter
Titles, links, and the 2021–2026 framing come from the accepted evidence register and checkpoint. This paragraph is an editorial synthesis, not an extract or an exhaustive account.
Open source record ↗08GovernCurrent inquiry
Human intent · Provenance · Authorization
Capability is not permission.
My current question is what happens between a person asking, a machine generating, and a system acting. I want the original intent, the constraints, and the decision to proceed to remain visible through that entire passage.
Behind this chapter
Present-day reflection, not a historical quotation or an assertion that earlier work anticipated modern AI.
09Build againCurrent build
Soval
Build the argument.
Soval brings the recurring questions into a practical working surface: keep human source distinct, make derivation visible, carry constraints, and leave authority with people.
Explore Soval ↗