Tim Negris has an old-fashioned talent that keeps becoming newly valuable: he can tell when a complicated technology needs plain words. In the early 1990s, while working at Oracle, he found two of them. “Thin client” described a model in which much of the computing work moved away from the user's machine and back toward a server. The phrase was short, argumentative, and useful. It gave Larry Ellison a weapon in the company's contest with desktop computing, and it anticipated an architecture that would later feel ordinary in the age of the cloud.
Negris did more than supply the label. He helped launch Oracle7 in 1993, during a difficult period for the company, and later called that launch his proudest professional achievement. His team let customers and recognizable industry voices carry the message. They even obtained a recorded comment from Bill Gates acknowledging Oracle's database leadership, an endorsement Negris recalled Gates later regretted. The detail belongs to a more theatrical era of enterprise software, when product battles came with characters large enough to bend the scenery.
Yet Negris has never sounded entirely persuaded by the theater. He has complained about ego wars and technology treated as religion, because users mostly want tools that help them finish their work and go home. That preference for use over spectacle is the straight line through a career that otherwise zigzags impressively: early artificial intelligence, databases, development tools, analytics, cloud architecture, machine learning operations, and now construction planning.
“Software is a tremendously powerful narcotic.”Tim Negris on what drives him
The data center came first
His first information-technology job was in a data center at 20. From there he moved into technical support, then training, then marketing. The sequence matters. Support teaches where systems break. Training teaches where explanations break. Marketing, done properly, notices both. Before the executive titles, Negris had learned to stand between a machine and the person expected to make sense of it.
He also worked around the first generation of artificial intelligence and was affiliated with John McCarthy, the Stanford computer scientist who named the field. Negris has recalled spending time with McCarthy and experimenting with LISP on Symbolics workstations. These were the years when AI's symbolic approach promised to represent knowledge through rules and logic, long before neural networks took over the public imagination. Having lived through multiple AI “mini-epochs” gives Negris a longer memory than the current boom usually rewards.
Four eras, one recurring job
His next landmark was relational data. Negris served as the first product manager for the database that became Microsoft SQL Server. He then moved through Sybase and Amdahl before Oracle recruited him. At Oracle, the server-centered idea needed a contrast with the increasingly powerful personal computer. “Thin client” did the job. Its survival is a small lesson in category design: a phrase sticks when it makes an invisible architecture easy to picture.
After Oracle came Hewlett-Packard and IBM. At IBM, he led the commercial introduction of data-mining and analytics software and helped take DB2 onto non-IBM hardware. The work broadened from marketing a database to explaining what companies could discover inside one. Data was becoming less a warehouse and more a source of decisions.
A résumé with sharp turns
The second half of Negris's career is less a march through household names than a tour of the software economy's restless middle. In the late 1990s he became president and chief executive of Omnis Software, the renamed Blyth Software, after joining in marketing and development. Omnis made rapid application development tools that could work across platforms, databases, and component systems. The company was smaller than the giants Negris had left, which forced a practical neutrality in the period's Java-versus-ActiveX quarrels.
He later served as vice president of marketing at Tivre Networks, spent several years consulting, and joined The SCO Group in 2005. Within months he moved from senior vice president of marketing to executive vice president of sales and marketing. A federal filing from that period supplies the sturdiest biographical footnote: Negris earned a B.S. from New York University.
The roles kept changing, but his management philosophy was consistent and pleasingly upside-down. He said he preferred to hire excellent people, discover what they could do, and shape jobs around them. “I create jobs to fit people, not vice versa.” It is the philosophy of someone who trusts range because his own career required so much of it: product judgment, technical fluency, positioning, writing, selling, and executive leadership.
There is a craftsman's streak outside the office too. Negris has listed photography and mountain biking as ways to relax, and he plays guitar. In a public note to a Massachusetts luthier, he admired the maker's insistence on shaping a new bridge saddle three times before it met the standard. The instruments he mentioned included Martin and Gibson acoustics. Technology careers are full of frictionless metaphors; a guitar is a useful corrective. It responds to small adjustments, but never becomes abstract.
Countries in which he has appeared as a featured speaker
Data center construction projects represented in customer consultations he discussed
AI meets concrete consequences
During the decade before MOCA Systems, Negris returned to artificial intelligence from the commercial side. His work included DataRobot and Rulex, then Spell, a platform for operationalizing deep learning that Reddit acquired in 2022. At Spell he promoted the idea of deep learning operations, the systems and discipline required to move a model from an experiment into reliable use. It was another category problem: the exciting algorithm receives attention, while the operational layer determines whether anyone can depend on it.
In July 2023, Negris joined MOCA Systems as chief marketing officer. Later company bylines and industry programs identified him as an executive vice president. MOCA combines consulting services with Touchplan, collaborative production-planning software for owners, designers, contractors, and trades. For a veteran of information systems, construction offers a bracing change of stakes. A database error can corrupt a report. A planning error on a large project can idle a crew, miss a delivery window, or send costs walking steadily uphill.
Negris joined with a clear brief. MOCA argued that its approach helped customers deliver complex projects with a success rate above 90 percent, against an industry in which failure to meet schedule, budget, or expectations was common. His response was pure marketer: “A great untold story is any marketer's dream.” The boast hides a serious preference. He wants a claim that can be measured.
Machine learning
Use historical project data to forecast bottlenecks, labor needs, procurement pressure, and schedule risk.
Deep learning
Interpret images and complex signals for progress tracking, recognition, translation, and inspection support.
Generative AI
Make project knowledge easier to retrieve, preserve expert judgment, and accelerate repetitive document work.
His construction writing divides AI into three working families. Machine learning looks for patterns in past project data and helps forecast what may happen next. Deep learning interprets complex visual and operational signals. Generative AI makes accumulated knowledge conversational and can speed repetitive planning and document work. The categories are less glamorous than a machine intelligence that does everything. They are also easier to assign, test, and supervise.
That practicality surfaced at ENR FutureTech in San Francisco in May 2026. Negris delivered a keynote called “The Three Faces of ConAI.” During the event, MOCA announced BoardScan, a Touchplan feature that photographs handwritten planning notes, interprets them, and turns them into sequenced digital tasks. The workflow is almost comically specific: snap, review, post. It also bridges a stubborn gap. Construction teams often plan together in physical rooms because sticky notes are quick and democratic; the resulting information still needs to enter a digital system. BoardScan attacks the transcription, not the meeting.
This is what Negris means when he says practical AI does tasks rather than jobs. The distinction cools the rhetoric without shrinking the opportunity. A narrow tool can save hours, preserve decisions, and create cleaner data for later analysis. The point is not to impersonate the superintendent. It is to keep the superintendent from typing a wall of notes into a schedule.
The cloud comes down to earth
Construction has also returned Negris to the physical machinery behind computing. He has consulted with customers connected to 200 data center construction projects and speaks about the densification problem: processors, storage, and networking equipment grow more powerful while occupying less space, concentrating heat and forcing cooling closer to the components. The artificial intelligence boom may sound weightless, but its appetite arrives as racks, pipes, power, water, skilled labor, and difficult schedules.
In his analysis, data center operators must balance the efficiency of liquid cooling against its capital cost, maintenance demands, and scarce expertise. Existing facilities face a harder retrofit decision than new builds. Demand is real, but so is uncertainty about how quickly generative AI will produce business value. It is a familiar Negris posture: interested in the new architecture, alert to the operational bill, unwilling to let enthusiasm cancel arithmetic.
That posture may be the most useful product of his long career. He watched symbolic AI give way to other schools. He saw database vendors fight holy wars. He helped turn a server architecture into a phrase, then watched the cloud absorb the argument. He led large-company launches and smaller-company reinventions. He has even named selling his Oracle shares, after a roughly tenfold rise during his time there, as his worst mistake. A person who admits that on the record is unlikely to confuse hindsight with prophecy.
“I create jobs to fit people, not vice versa.”A management rule built for versatile teams
Negris's present work is not a departure from the early chapters. It is their physical sequel. SQL Server organized enterprise information. Thin-client computing moved work across a network. Analytics searched data for patterns. DLOps made machine learning repeatable. Construction AI is now asked to read the plan, notice the risk, preserve the expert's knowledge, and help a crew keep its promise for Friday.
The striking part is not that he has repeatedly been near the next thing. Technology careers are generous with next things. It is that he keeps asking the same unfashionable question of each one: what, precisely, can people do with it? Sometimes the answer needs an operating system. Sometimes it needs a photograph of a planning board. And sometimes it needs two words that everyone remembers.