Andrew Feinberg has the sort of résumé that makes a neat category feel slightly underdressed. He is president and CEO of BostonGene, a biotechnology company built around AI and multimodal biological data. He is also chairman and CEO of Netcracker Technology, the enterprise software business whose systems sit deep inside telecom operations. One job deals with tumor biology and drug development. The other deals with networks, billing, operations and customer journeys. At first glance, the overlap looks like a clerical error.
Look longer and a single theme emerges. Feinberg has spent his career around systems that create more information than any one person can sensibly hold in mind. His work is about arranging that information so somebody can make a better decision. The nouns change. The operating verb does not.
That continuity matters because corporate biographies tend to sand careers into a smooth sequence of titles. Feinberg's is more interesting as a record of transferred judgment. He began in strategy consulting, including work connected to telecommunications, cable, software and hardware. He joined Netcracker in 1998 and became part of its senior leadership. More than a quarter century later, he remains at the company, now as chairman and CEO.
The long apprenticeship in complexity
Telecom software rarely enjoys the glamour attached to consumer apps. It does, however, offer an exacting education in scale. A communications provider must connect services, customers, devices, networks, orders and bills while the underlying technology keeps changing. A clever feature is not enough. The platform must work inside a living system with old components, new demands and very little patience for downtime.
Feinberg's public comments from that world return to customer-centricity. He has argued that digital transformation cannot be reduced to technology alone; it also needs operational, cultural and business support. The statement is practical to the point of impoliteness. Buying software is easy. Changing how an organization behaves around it is the expensive part.
That view also explains his preference for platforms over isolated tools. Platforms can bring scattered signals into one working context. They can support a chain of decisions instead of producing a dazzling result that has nowhere to go. At Netcracker, the context is a telecom operator. At BostonGene, the context is biology, research and therapeutic development. The second domain carries its own expertise and constraints, but the management problem feels familiar: fragmented data, complicated users and consequences that cannot be waved away with a demo.
The second room
BostonGene was founded in 2015. Feinberg is publicly identified as a co-founder and investor, and he became president and CEO in 2019. The company's proposition starts with a complaint about fragmentation. Genomic data can describe one layer. Transcriptomic data offers another. Immune, spatial and clinical signals add further context. Examined separately, each can be useful. Integrated, they may reveal a pattern that the pieces do not show alone.
Feinberg reaches for a musical metaphor to make the point. “You cannot identify what a symphony is playing by just looking at the instruments,” he said in a 2025 life-sciences interview. It is a good line because it refuses the standard worship of data volume. Owning a larger inventory of instruments does not guarantee music. Arrangement is the work.
The recurring Feinberg operating loop
signals
context
decision
BostonGene's platform combines multiple forms of biological data with AI models and laboratory capabilities. The company sells into two linked settings: biopharma research and clinical development, where teams are trying to select indications, biomarkers and trial populations; and precision care, where molecular context can inform clinical decision-making. Feinberg's language has increasingly shifted toward “decision intelligence.” The phrase is corporate, but the distinction underneath it is sharp. Analytics describes. Decision intelligence is expected to change what happens next.
The market rewarded the ambition. In April 2022, BostonGene announced a $150 million Series B led by NEC Corporation, with participation from Impact Investment Capital and Japan Industrial Partners. The financing was intended for research, clinical partnerships and international expansion. Feinberg described the opportunity as the rare combination of large economic value and even larger social impact. He sounded excited, and a little wary of the scale of the promise.
The NEC connection was not a decorative investor logo. NEC acquired Netcracker in 2008, and in 2023 BostonGene, NEC and Japan Industrial Partners formed BostonGene Japan. That joint venture turned corporate adjacency into an operating bridge: BostonGene's platform, NEC's technology reach and a local investment partner organized around expansion in Japan.
A partnership map, not a lone-genius tale
BostonGene's public partner list reads like a map of how modern biotechnology actually moves: pharmaceutical companies, academic medical centers, research networks and technology partners. It includes organizations such as AstraZeneca, Johnson & Johnson, Takeda, Dana-Farber Cancer Institute, Mass General Hospital, Johns Hopkins School of Medicine and universities in Japan. The arrangement is a clue to Feinberg's operating style. The platform may be centralized; the knowledge is not.
His own public posts habitually distribute credit to teams, collaborators and customers. That may be ordinary executive etiquette, but it is consistent with the underlying model. Multimodal work depends on multiple disciplines. Software engineers, biologists, researchers, clinicians and commercial partners must agree on interfaces, evidence and what counts as a useful output. The charismatic solo act would be a poor organizational design.
This also makes Feinberg's double role less eccentric than it appears. He is not moonlighting between unrelated products. He sits at the junction of two ecosystems connected through NEC and a shared appetite for AI infrastructure. One works across massive communications networks. The other models biological systems. Both require integration, governance and trust before intelligence can operate at scale.
The intelligence dividend needs a chaperone
By 2026, Feinberg's argument had moved beyond the usefulness of analytics to the conditions under which AI should be trusted. At Mobile World Congress Barcelona, he delivered a keynote on governing AI at scale. The announced focus was transparency, rigor and human oversight across high-impact industries. It was a fitting stage: a global telecom gathering hosting an executive who planned to carry lessons from oncology AI back into the wider technology ecosystem.
Governance can sound like the gray cardigan at the innovation party. Feinberg presents it as infrastructure. If a model affects an expensive trial, a complex network or any consequential workflow, the organization needs to know what the system is optimizing, where its evidence comes from and when a person must intervene. Scale without those answers does not produce an intelligence dividend. It produces a larger mystery.
BostonGene's 2026 announcements pushed the same theme through collaborations and recognition. The company announced work with AstraZeneca and Johnson & Johnson, continued expanding academic partnerships in Japan, and was named Overall AI-based Analytics Company of the Year by the AI Breakthrough Awards. Awards are not scientific validation, and partnerships are not outcomes. They are, however, evidence of the route Feinberg has chosen: place the platform inside serious institutional work, where claims encounter data, timelines and demanding counterparties.
What builders can steal
The first lesson is to define a career by the problem you solve, not only by the industry printed on the badge. Feinberg moved from consulting to telecom software to biotechnology without pretending those disciplines were interchangeable. The portable asset was a way of seeing: find the fragmented signals, build the integration layer, and connect the result to a decision.
The second is that personalization is an architecture, not a slogan. In telecom, it requires current customer and network context. In biology, it requires multiple data modalities and domain expertise. In both cases, the generic output is the easy one. The difficult product remembers who the decision is for.
The third is less fashionable: remain long enough to understand the plumbing. Feinberg joined Netcracker in 1998. That tenure spans several generations of telecom technology and multiple cycles of corporate enthusiasm. It is difficult to speak convincingly about transformation if you have never stayed to watch one finish.
A fourth lesson hides in the way he talks about impact. Feinberg rarely separates the technical system from the institution using it. His public remarks pair models with workflows, platforms with partnerships, and intelligence with oversight. For a builder, this is a useful correction. Accuracy on a benchmark may open a door, but adoption depends on training, incentives, interfaces, evidence and responsibility. The product must survive contact with the organization. Feinberg learned that in telecom, where a platform crosses departments and legacy systems. BostonGene applies the same patience to a setting where disciplines bring different vocabularies and standards of proof.
There is an appealing symmetry to his current chapter. The telecom executive who learned to organize data at industrial scale now argues that AI in biotechnology needs biological grounding and human accountability. The biotech executive returns to telecom's global stage to discuss what rigorous AI governance looks like. Each room corrects the excesses of the other.
Feinberg's career is therefore not a story about a man collecting impressive titles. It is a long experiment in making complexity useful. The experiment has produced companies, financing rounds, partnerships and plenty of acronyms. Its cleanest measure remains humbler: when the data arrives, can someone see the next move?