Abir Ray
CEO, Expression
Abir Ray leads efforts to scale Expression’s VOR Platform and ARGUS multi-agent reinforcement learning system into production deployments across defense and intelligence community programs, while continuing to deliver on major prime contracts including the NTIA Federal Spectrum Data System and DISA SIS and MSP contracts.
He also co-authored “Spectrum Aware Systems Engineering: AI, Digital Twins and MBSE for Electromagnetic Operations” (Wiley-IEEE Press) with Dr. John Stine of MITRE. The book bridges urgent defense mission needs with rigorous systems engineering methods agencies can adopt today.
“The biggest unlock for federal missions is not larger models — it is better systems engineering around the models we already have,” Ray said. “Our focus at Expression is putting agentic AI inside the operational fabric of government in ways that pass audit, hold up under adversarial conditions and give analysts back the time to do the work only humans can do.”
Why Watch
In 2026, Ray’s team is heavily focused on operationalizing agentic AI inside high-assurance government environments where accuracy, explainability and auditability are non-negotiable. They are advancing VOR’s data, workflow and agent modules alongside ARGUS’s multi-agent reinforcement learning stack to support spectrum operations, intelligence analysis and cyber decision support. Several deployments will move from pilot to program of record this year, he said.
Ray’s team helps customers operate more efficiently by replacing brittle, manual analyst workflows with closed-loop agentic systems that handle high-volume, low-judgment work and surface the decisions humans actually need to make. On programs like NTIA FSDS and DCSA’s counterintelligence analytics work, that turns weeks of spectrum or CI triage into hours, with full provenance and human-in-the-loop control at every step, he said. The platforms are architected so government teams retain ownership of the data, the models and the decisions, which is what makes the efficiency durable rather than dependent on a single vendor, he added.
“Abir is one of the most technically credible founder-operators we work with in the federal market,” said Devin Talbott, founder and CEO of Enlightenment Capital. “He has built Expression into a platform company with genuine IP in agentic AI and spectrum operations, and the discipline to deliver on some of the most demanding programs in the Department of Defense and intelligence community. What sets him apart is the combination of deep research instinct and operator pragmatism.”
Fun fact: Ray is a doctoral candidate at the University of Tennessee researching Lyapunov-based stability boundaries for multi-agent reinforcement learning on USRP B210 software-defined radios. His ideal weekend is a draft chapter in the morning followed by hanging out with friends in the afternoon.