Lazarus Systems is an independent AI laboratory. We combine experiments with software development: testing models, designing integrations and preparing systems for everyday use.
SEE WHAT WE WORK ON
We shapethe future
with AI.
We build AI agents, automate workflows and connect models to data. We also build the infrastructure you need to run them.
Models and data. Experiments and deployments.
SCROLL TO EXPLOREAmbitious ideas.
Real solutions.
Start with the task
We define what the system needs to do and how we will assess its output. Then we compare approaches and develop the one we choose.
Choose where it runs
We design systems for local, cloud or hybrid environments. The choice depends on your data, costs and access requirements.
Technology
for business growth.
We can work on a specific integration or an entire system, from preparing the data to deploying and monitoring the model.
AI agents
REASONING / ACTIONWe build agents that use tools and carry out tasks in existing systems. We design their memory, integrations and the points where a person reviews and approves their work.
Workflows & integrations
CONNECT / AUTOMATEWe connect applications and design the workflows between them. We decide which steps need a model, which need conventional code and which need a human decision.
Data analytics
EXPLORE / UNDERSTANDWe organise data, examine relationships and prepare analyses. We build tools that let you return to the data and check how a conclusion was reached.
Data systems & RAG
COLLECT / RETRIEVEWe build data pipelines and semantic search. We connect models to documents so they can answer questions using your organisation’s knowledge.
AI infrastructure
BUILD / OPERATEWe set up environments for models and agents, including deployment, monitoring, evaluation, permissions and cost control.
Local AI
PRIVATE / LOCALWe run models on your organisation’s infrastructure. We select models suited to your hardware and the tasks they need to perform without external model APIs.
Intelligence
at a new scale.
We connect AI, data and infrastructure to support your organisation’s growth.
Expanding
the horizons of AI.
Experiments help us assess where a model is useful and where its limits lie. Our work covers four areas.
Agent collaboration
We explore task allocation, context sharing and human oversight of agents.
Document retrieval
We test how source selection and context preparation affect a model’s answers.
Locally hosted models
We compare response quality and speed with limited hardware resources.
Processing incoming data
We explore how to detect changes in data and use them to trigger the next steps in a workflow.
From inside
the laboratory.
Notes on models, data and the decisions behind AI systems.
Artificial Analysis results, Anthropic pricing and API changes. Lazarus Systems on choosing settings and assessing the cost of an AI agent’s work.
Read the articleLet’s shape
what comes next.
Describe the task, workflow or system you want to build. If you already have data or a prototype, tell us where you are in the process.