Decision advantage
is a weapon system.
We give commanders, analysts, and program managers assessments they can act on fast and defend under scrutiny. When a question keeps coming back, we build the tools that keep answering it.
The Problem
You don't have a data problem. You have a synthesis problem.
Your analysts are drowning in feeds, reports, and dashboards. The data is there. The synthesis layer that turns it into a defensible decision at OODA speed is what's missing. We supply it.
Meanwhile, every AI vendor in your inbox is promising "revolutionary" capabilities with demos that look great on commercial data and fall apart on yours. You've seen this movie before.
AI pilots that never reach production
Impressive demos, zero operational impact. The CDAO has identified scaling from pilot to production as the #1 barrier to DoD AI adoption.
Black-box outputs you can't brief upward
"The algorithm said so" does not survive a Congressional hearing, a GAO audit, or an IG investigation. Commanders need to show their work.
Consulting decks gathering dust
You paid for strategy. You got PowerPoint. The recommendations were outdated before the ink dried because the environment changed and the analysis couldn't.
Vendor lock-in on your own data
Proprietary platforms where government data goes in and doesn't come out. No API, no interoperability, no data sovereignty.
A chokepoint on the map
The research under military machine vision runs through one community. Its new work comes from China, eight to one.
We measured which research communities the rest of science leans on hardest. At the top of the list is the core of deep-learning object detection, the community that holds ResNet and Faster R-CNN. 423 other communities take at least 5 percent of their outside citations from it. Ship detection in radar imagery takes 47 percent.
Most of what those fields cite is the older foundation, published before 2020 and open to anyone. The new work is another matter. Of the papers from 2022 onward that those 423 fields cite, Chinese institutions wrote 47 percent. American institutions wrote 6.
Reading that research is sound practice. What the ratio tells you is where the next detection method is likely to come from, and how little of it is now written at home, in work that sits under ISR and autonomy.
423
communities depend on it, more than on any other in the map
47 / 6
percent of its cited new work from Chinese and American institutions
Share of each field's outside citations drawn from the detection core
- Ship detection in SAR imagery47%
- Infrared and visible pedestrian detection34%
- Prohibited items in X-ray baggage scans32%
- Infrared small target detection23%
- Multi-object tracking22%
- Weapon detection in surveillance video21%
- 3D object detection for autonomous driving18%
- Drone detection and classification13%
Every mission rests on communities like this one. We can map the ones under yours and show who writes them.
Request a dependency mapMission Areas
Organized around your mission
Competitive Assessment & Strategic Intelligence
Strategic assessment across military, economic, and technological domains. We bring analytical frameworks that work across open sources and long time horizons to surface competitive asymmetries and generate testable hypotheses.
- Multi-domain competitive assessment (military, economic, technological)
- Victory hypothesis generation and ACH-grounded COA development
- Assumption surfacing, red team analysis, and war-gaming support
- Long-horizon strategic trend analysis with quantified confidence
Technology Scouting & S&T Assessment
Mapping the global S&T literature through its 1.7 billion citation links, alongside patents and emerging capabilities. Identifying who is working on what, where the breakthroughs are coming from, and what they mean for national security.
- Global bibliometric network analysis and research mapping
- Emerging technology identification and capability forecasting
- Foreign S&T assessment and research security evaluation
- Innovation pipeline discovery across adversary ecosystems
Decision Architectures for Complex Operations
Structured decision support for coalition-grade, multi-stakeholder problems. Transforming ambiguous, high-stakes situations into actionable frameworks with explicit assumptions, causal models, and auditable reasoning chains.
- Sensor-to-decision pipeline acceleration
- Bayesian confidence calibration and causal analysis
- Multi-perspective synthesis across intelligence disciplines
- Decision audit trails for Congressional and IG oversight
Net assessments with verbatim, checked citations and dated predictions that get scored, built to sit inside your existing workflows. Today that work runs on unclassified sources.
Past Performance
We've done this work before.
Our team's defense work goes back more than a decade, from Melissa Flagg's years overseeing DoD science and technology programs to the science maps and S&T intelligence we've built for DARPA and CSET. Syntheos was the prime on the DARPA, CSET and Noble Reach work.
DARPA
An analytic study of the diffusion, foreign uptake, and commercial potential of the research outputs from a DSO-funded research program (award HR0011263E005, October 2025 through July 2028).
Defense Advanced Research Projects Agency
CSET
Built science mapping, automated text annotation and classification systems, and S&T growth prediction models for analyzing emerging technology trajectories across global research ecosystems.
Center for Security and Emerging Technology
Marshall Foundation
Built a research console over the Foundation's digital archive of Office of Net Assessment materials. Cited archive content and model inference are distinguished throughout, so researchers can think with the corpus without confusing what it says with what the model extrapolates.
The Andrew W. Marshall Foundation
Noble Reach
Built systems for automated extraction and classification of S&T problem statements, paired with regional expertise matching to connect mission needs with the researchers and institutions best positioned to address them.
Noble Reach Foundation
References available on request.
DoD AI Alignment
Built for the standards the CDAO is setting.
The Chief Digital and AI Officer has been explicit. DoD wants AI that scales past pilots and models that are explainable. Those are the standards we hold our work to.
Built for operations
When an engagement calls for a tool, it is designed for sustained operational use, with model governance and continuous monitoring from the start.
Explainable & Auditable
Every claim carries its citations, assumptions are stated, and our analysts check the key judgments against the sources by hand. Built for leaders who answer to Congress, GAO, and the Inspector General.
Human-Machine Teaming
AI augments analysts and decision-makers, and human judgment stays in the loop. The reasoning stays visible. Your data stays yours.
Plain about security
Syntheos does unclassified work today. NIST SP 800-171, DFARS 7012 and CMMC Level 2 are in progress, and the table below says so. When a tool has to work with your own sources, we build it against synthetic data shaped like them, and your team connects the real sources inside your boundary.
Security & Compliance
Compliance posture
NIST SP 800-171
In ProgressCUI protection controls
DFARS 252.204-7012
In ProgressSafeguarding covered defense information
CMMC 2.0 Level 2
In ProgressThird-party assessment
Responsible AI
By DesignDoD AI Ethical Principles aligned
Compliance documentation available upon request as certifications are completed.
How to Acquire
Acquisition pathways
NAICS Codes
Product Service Codes
Contract Vehicles & Mechanisms
- Direct award and competitive task orders
- Other Transaction Authorities (OTAs)
- Broad Agency Announcements (BAA)
- Open to Commercial Solutions Openings (CSOs)
- Open as subcontractor on major IDIQs
- GSA Schedule (in progress)
Registration
- UEI
- K9WYGF15HZV8
- CAGE
- 12TM0
- Business size
- Small business
- SAM.gov
- Registered and active
Everything a teaming agreement or a sources-sought response needs is on this page. For anything else, write to defense@syntheos.io.
Caleb Smith
Founder & CEO
Published researcher in bibliometrics and research intelligence. Caleb is a partner at SciTech Strategies and Director of Research Intelligence & Analytics at the University of Michigan Medical School, where he builds analytical frameworks for mapping global S&T ecosystems, detecting scientific innovation, and predicting research trajectories.
At Syntheos he leads the work for clients including DARPA, CSET, and the Marshall Foundation, and he built Deepfield, Syntheos's strategic assessment platform.
Partner, SciTech Strategies • Director, Research Intelligence & Analytics, University of Michigan Medical School • DARPA, CSET, Marshall Foundation
Melissa Flagg
Senior Advisor
Former U.S. Deputy Assistant Secretary of Defense for Research, where she oversaw DoD science and technology programs from basic research through advanced technology development. Dr. Flagg is Executive Director of the Andrew W. Marshall Foundation, a Syntheos client, and a Senior Advisor to Georgetown University's Center for Security and Emerging Technology. Her prior roles span the Office of Naval Research, the Department of State, and the Office of the Secretary of Defense for Acquisition, Technology & Logistics. Before joining the Pentagon as Deputy Assistant Secretary, she was a senior program officer in the MacArthur Foundation's fellows program, focused on science, technology, engineering and mathematics. She holds a Ph.D. in Pharmaceutical Chemistry and has served on the National Academy of Sciences Air Force Studies Board.
Kevin Boyack
Chief Technology Officer
Nearly three decades of expertise in scientometrics, science mapping, and research evaluation. Dr. Boyack serves as President of SciTech Strategies, where he builds global models of scientific literature and predictive analytics for strategic planning and competitive intelligence. Previously, he spent 17 years at Sandia National Laboratories across combustion research, transport processes, and socio-economic war gaming. He holds a Ph.D. in Chemical Engineering from Brigham Young University and has authored over 80 publications on science mapping and citation analysis.
Richard Klavans
Chief Science Officer
Over 30 years of experience in science and technology strategy, with a focus on mapping the global structure of scientific knowledge. Dr. Klavans founded SciTech Strategies, building one of the first global models of scientific literature for strategic planning and competitive intelligence. He has advised Fortune 100 companies, government agencies, and leading academic institutions on R&D portfolio management. A SCIP Fellow and former President of the Society of Competitive Intelligence Professionals, he has co-authored over 90 peer-reviewed publications on research evaluation and science policy.
Initiate Contact
Ready to discuss your mission requirements?
Tell us about the decision problem you're trying to solve. We'll set up an unclassified briefing with our principals to assess fit and discuss acquisition pathways.
Direct Contact
Defense programs: defense@syntheos.io
Contracting officers: contracts@syntheos.io
Everything else: info@syntheos.io
Classification
Syntheos does unclassified work today. Please don't send classified or controlled information through this form or to any of the addresses above.