Research that shapes what you build next
Applied research in AI, machine learning and IoT security.
Overview
Research from the teams who design, build and run systems for our clients, so every finding connects to something we deliver.
We are open to collaborating with industry, academic institutions and PhD aspirants.
Our research
What we research
Current work in language models, and completed research in recommendation systems, machine learning and IoT security.
Large language models (LLMs)
Applying language models to business work: answers grounded in company knowledge, document understanding and safe, supervised automation.
Data & AIRecommendation systems with LLMs
Combining language models with user and product signals to make relevant, explainable recommendations.
Data & AIDecision tree models
Interpretable machine-learning models for classification and everyday business decisions.
Data & AIIoT security
Protecting connected devices, their data and their networks, from the sensor to the platform.
Cybersecurity · IoTIoT security with blockchain
Device identity and tamper-evident records for IoT using blockchain.
Cybersecurity · IoT“Our research exists so our clients' next system is safer and smarter than their last, especially in AI and IoT security.”
Research tracks
A research track for every service
Every piece of research belongs to a service, so insight always connects to something we deliver.
Perspectives
What we are writing about
Our published series, starting with customer experience.
Technology radar · Edition 1 · Q4 2026
What our engineers adopt, trial, assess and hold
A living view of the technologies we recommend to clients, refreshed every quarter.
Adopt
Our default choice for new client work.
- Java 21 + Spring BootProven backbone for multi-tenant business platforms.
- React and Next.jsFast, accessible web front ends with server rendering.
- FlutterOne codebase for Android and iOS field and staff apps.
- PostgreSQLReliable relational core for transactional products.
- Docker and CI/CD pipelinesRepeatable builds and releases on every project.
Trial
In production on selected projects; expanding with care.
- Retrieval-augmented generation (RAG)AI answers grounded in a client's own documents.
- Kubernetes for larger workloadsWorth it once services and traffic justify the overhead.
- Infrastructure as code (Terraform)Reviewable, repeatable cloud environments.
Assess
Being evaluated in pilots and internal builds.
- Agentic AI workflowsMulti-step automation with human approval points.
- Edge analytics for IoTDecisions on the shop floor without a round trip to the cloud.
- Blockchain for traceabilityTamper-evident records across supply and audit chains.
Hold
Not recommended for new work.
- New builds on unsupported frameworksSecurity and hiring costs rise every year.
- Chatbots without connected dataThey answer questions the business cannot act on.
Research briefs
Briefs in preparation
Short, practical briefs written by ATOCONN specialists, drawn from our research and delivery work.
Securing connected operations: IoT, identity and tamper-evident records
How manufacturers and audit-led businesses protect device data from sensor to report.
Cybersecurity · IoTDesigning journeys customers return to
The CX series in one practical brief: mapping, connected data and measurable experience goals.
CX TransformationAI readiness for growing businesses
What to have in place — data, ownership and use cases — before the first AI project.
Data & AICollaborate with ATOCONN
Have a research project in mind?
Tell us about your project or the area you would like to explore together.
- IndustryJoint research on real business data and systems
- Academic institutionsProjects, guidance and shared research work
- PhD aspirantsResearch questions grounded in live products

