Applied Software R&D
We research, prototype, and engineer specialized software systems where the solution—or even the right technical approach—is not yet obvious.
Our work combines software engineering, applied AI, automation, and deep engagement with specialized problem domains to turn promising ideas and difficult technical questions into evaluated prototypes, practical architectures, and working systems.
Some technical problems begin with a specification. Others begin with an unanswered question.
Applied Software R&D is useful when important technical uncertainty remains—when the problem is understood but the solution is not, when an emerging technology looks promising but its value has not been demonstrated, or when the fastest path to an answer is to investigate, experiment, and build.
You understand the problem or opportunity, but meaningful questions remain about feasibility, architecture, technology, or the right technical approach.
We use focused research and engineering to explore the solution space, test important assumptions, and develop evidence before difficult-to-reverse technical commitments are made.
Your data, workflows, constraints, or technical environment make off-the-shelf approaches a poor fit.
We engage with the realities of the problem domain and use that understanding to shape the abstractions, architecture, evaluation criteria, and technologies appropriate to the problem.
AI, agentic systems, automation, or another emerging technology may create an opportunity, but adopting the technology is not itself the objective.
We investigate where it provides meaningful advantage, where conventional engineering remains the better choice, and how the two can be combined when a hybrid approach is more effective.
Some questions cannot be answered adequately through analysis alone.
A focused experiment, proof of concept, or prototype can expose constraints, challenge assumptions, evaluate competing approaches, and generate the evidence needed to determine what should come next. The prototype does not necessarily need to become the product; sometimes its most important job is helping determine what the product should be.
Our R&D work spans four areas of technical capability. Depending on the problem, an engagement may draw from one area or combine several.
We research and engineer agentic systems, domain-specific AI applications, model integrations, structured AI workflows, evaluation mechanisms, and human-guided automation.
Our approach treats AI as one engineering capability among several. Where appropriate, we combine probabilistic model capabilities with conventional software, deterministic analysis, and explicit workflow control.
We develop tools and systems that improve how software is built, tested, analyzed, and maintained.
Our work includes developer tooling, static analysis, engineering intelligence, workflow automation, testing infrastructure, and specialized internal systems designed to reduce engineering friction, improve consistency, and make important development activities easier to inspect and verify.
We investigate software approaches for understanding network activity, security-relevant data, and executable software.
Our work includes network analytics, packet-data systems, executable static analysis, specialized representations of network activity, synthetic cyber data, and AI-assisted security research.
We research and engineer software for complex creative and production workflows where automation, generative technologies, structured information, and human judgment must work together.
Our work includes AI-enabled production systems, workflow orchestration, asset pipelines, generative-media infrastructure, and specialized creative applications.
Applied Software R&D can begin at different points in the life of an idea or system. An engagement may focus on a single stage or combine several as the work develops.
Is the idea viable, and what approach should we take?
We research alternatives, evaluate technologies and models, conduct focused experiments, identify constraints, and test consequential assumptions.
The objective is evidence: enough to make a better-informed decision about whether to proceed, change direction, simplify the approach, or stop.
Can we demonstrate the idea and determine what the real system should be?
We build focused experiments, proofs of concept, and prototypes to explore system behavior, refine requirements, evaluate architectural boundaries, and expose constraints that may not be apparent through analysis alone.
The prototype is not necessarily the product. It can be an instrument for learning what the product should be.
Can we build this and make it work within our environment?
We design and implement the software, automation, integrations, internal tooling, APIs, services, and other components required to turn an evaluated approach into a working system.
Can we depend on it, operate it, and ultimately own it?
Where R&D is ready to move toward operational use, we focus on the engineering required to make the resulting system more dependable and transferable, including testing, observability, deployment considerations, failure modes, documentation, and operating procedures.
The objective is to build capability with clients rather than unnecessary dependency on the consultancy.
Engage us at any stage. We can investigate an uncertain technical opportunity, demonstrate a viable approach, engineer the resulting system, or help transition existing R&D toward operational software.
Selected research and engineering projects exploring difficult technical problems through investigation, prototyping, and software engineering.
Characterizes the semantic nature of software changes > to help reviewers identify where deeper analysis may be warranted.
Analyzes characteristics of software that create friction in understanding, reasoning about, and safely modifying it.
Represents and preserves structured provenance describing the origin, lineage, and transformation of AI-generated and transformed artifacts.
Translates an analyst's investigative intent into an executable strategy for analyzing captured network traffic.
Statically analyzes executable artifacts to locate where selected imported functions are actually called, creating a foundation for deeper analysis of the behavior surrounding those calls.
Not every useful technical investigation needs to become a large system or long-running R&D program.
The Engineering Lab collects focused tools, experiments, and exploratory work addressing specific engineering problems and questions.
Automating consistency checks between authoritative Enum definitions, configuration declarations, and source-code references to detect integration errors before they become runtime problems.
Providing deterministic, configurable responses that emulate LLM-provider APIs so applications can be developed and tested without depending on a live inference service.
Exploring developer tooling for making merge-request preparation, repository state capture, mandated rebases, and related integration activities more consistent and diagnosable.
The consultancy's technical work is led by Jonathan Jarrett, a senior software engineer and technical leader working across software architecture, applied AI, developer engineering, cybersecurity and network analysis, and specialized software R&D.
Jonathan remains directly involved in technical investigation, architecture, experimentation, and engineering. His work focuses particularly on problems where important technical questions remain unresolved and where working software can help determine the right path forward.
His approach combines deep engagement with the problem domain, explicit evaluation of technical tradeoffs, and pragmatic selection of technologies. Emerging technologies such as AI and agentic systems are considered alongside conventional software engineering, automation, deterministic analysis, and hybrid approaches according to the needs of the problem.
Through the consultancy, organizations can engage this senior technical capability for focused R&D, architecture, prototyping, specialized engineering, and collaborative technical initiatives through a B2B relationship.
You don't need a complete specification—or even certainty about the right technical approach—to start a conversation.
If you're evaluating a promising technical idea, confronting a problem that doesn't fit a generic solution, exploring where AI or automation could provide meaningful value, or trying to determine what should come after an experiment or prototype, we can begin with the questions that remain unresolved.
The first step is understanding the problem, the context around it, and what you need to learn or accomplish. From there, we can determine whether focused research, experimentation, prototyping, engineering, or another approach makes sense.
Start a Conversation