Two founders. Nine companies. Four exits.
BrightMeld is run by its two founders — no bench of juniors, no offshore delivery arm. Between us: nine companies founded or led, exits to EarthLink, Symantec, and AnswerThink, and technology we built acquired by the U.S. Department of Homeland Security. We've been building enterprise software since before "the cloud" had a name, and we've built together before — successfully.
Marc Epstein
Managing Partner & Co-Founder LinkedIn ↗Marc has built software at every layer of the stack — and built the teams and companies that shipped it.
He started where the stack starts: operating systems, at AT&T Bell Laboratories and the Unix Systems Laboratory, then led product management for the UnixWare Technology Group as Unix itself went to market. He ran a $100M networking division as GM/VP at Novell. As CTO of Quarterdeck, a public PC-software company later acquired by Symantec, he directed all R&D and productization. He co-founded CenterBeam — helping raise over $180M in venture funding across his career — which was acquired by EarthLink, and co-founded Data Mining International, whose fraud-detection technology was acquired by the U.S. Department of Homeland Security. At MyExposome, which he founded, he patented the company's technology himself and personally built its data analytics and reporting systems. He holds four U.S. patents spanning data, network, and software.
At BrightMeld, Marc designed and implemented the pipeline itself — the nine-gate methodology that turns AI construction into contract-driven, audited delivery. Four decades of building and leading distributed engineering teams taught him what the job of shipping software actually is. The pipeline is that job, encoded.
Operating systems, networks, applications, enterprise systems — the whole stack. And now, the thing that builds it.
Mort Goldman
Managing Partner & Co-Founder LinkedIn ↗Mort delivers enterprise systems.
That's been the job for four decades and four complete turns of the technology, and it started at Computer Corporation of America — the 1980s research hub behind Model 204 — where he led application development on Model 204 for government and commercial clients while working on Multibase, a distributed heterogeneous database built on a functional data model. Getting incompatible systems to answer one question coherently was a research problem then. It's still most of what enterprise software gets wrong.
He was Lead Architect in Xerox's Advanced Information Technology Division, working on distributed database and workflow systems, then VP Technology at Information Technologists. In 1993 he founded ICON Solutions and grew it to $25M in revenue and 120 people before AnswerThink (NASDAQ: HCKT) acquired it. He went on to found Fortera, which insured enterprise IT projects against non-completion — a business that only works if you can look at a build early and tell whether it's going to fail, and why. He served as CTO of Genome Profiling, and today leads enterprise architecture at Quench USA, where he has held VP roles across IT, business intelligence, and architecture and has been responsible for the design and implementation of its core transactional and analytical systems.
Different industries, different decades, same role: the one accountable for shipping.
There's no bench of juniors here because there doesn't need to be one. AI does the construction, under nine human gates and an independent review designed to distrust it. Two people who have shipped this kind of system for four decades decide what's correct. That's not a smaller version of a consultancy — it's a different shape.
BrightMeld is new. We aren't.
This firm doesn't have a long client list yet, and we'd rather price that honestly than imply otherwise. So being early is priced: the Spark assessment costs you nothing — you keep the full specifications and estimate whether or not you go further — and your first build is fixed price, with overruns on us, not you. Those terms aren't bravado. One of us founded and ran a company that underwrote enterprise IT projects against non-completion; we know exactly what this risk costs, because we used to sell insurance against it. And because you own the code outright from day one, your downside is bounded no matter how we perform.
Two today, and that's the shape of the firm rather than the ceiling. We'll add senior people who own what they build — not layers to bill through.
We've shipped together before.
BrightMeld isn't our first collaboration. At Data Mining International, Marc built and led the company while Mort architected and directed product development of its fraud-detection technology for international trade.
Book a free Spark assessment