01
Start from the running process.
We begin with what the operation is trying to do, which signals and systems are involved, where the current path slows down or loses state, and how the team knows the result was successful.
Company
We work on the path between information and outcome.
Amalgus works with customers to build and improve software products and operating systems that cross data, context, AI, workflows, infrastructure, and people. We make the necessary state available, put decisions inside clear authority, carry them into execution, and expose evidence of the result.
Our role is usually to trace the end-to-end path, isolate the real constraint, build the missing pieces, instrument the system, and stay with it long enough to see how it behaves in use.
How we approach the work
We stay close to the product, the measurements, and the people who use and operate it.
02
Keep state close to the decision that needs it.
History, asset identity, process state, permissions, and relationships only help when they are available at the right place and are fresh enough for the operating decision.
03
Follow every decision through execution.
A recommendation, model output, or rule is only useful once the next step has been carried out and the resulting state has been checked.
04
Design for the bad hour, not the average hour.
Bursts, retries, queue growth, shared resources, stale state, partial failure, and recovery determine whether a system remains usable when the operation is under pressure.
Engineering depth
The work crosses disciplines because the outcome crosses systems.
We bring the disciplines needed for the operating path and keep ownership clear at the boundaries. That is where latency, state, authority, retries, and failure behavior tend to become someone else's problem.
Sensing + perception
LiDAR, video, telemetry, tracking, classification, anomaly, retrieval, prediction, and confidence.
Data + context
Streaming and batch data, knowledge and graph layers, history, state, lineage, freshness, identity, and semantic relationships.
Decision + action
Rules, models, optimization, recommendations, agents, permissions, approvals, workflows, APIs, machines, and operators.
Execution + scale
Distributed runtime, edge/cloud placement, queueing, scheduling, backpressure, consistency, observability, recovery, and capacity.
Trust + lifecycle
Confidential compute, HSM/root of trust, secure boot, SBOM/provenance, release evidence, hardware/firmware, and sustaining.
Experience
The focus comes from years of working across products, platforms, data, AI, infrastructure, and operations.
Prior work includes product and developer platforms, payments, large-scale data environments, recommendation systems, RAG, governed agent workflows, wireless and edge infrastructure, LiDAR + AI perception, and operational intelligence. The useful common thread is the end-to-end path: enough state, available in time, leading to an action the operation can trust.
Careers
Work on systems where timing and state matter.
We are interested in engineers who can go deep in their discipline and still follow a problem across interfaces, failure modes, workload behavior, and the real operating result.