Telecom · home networking
WiFi log intelligence with RAG
Vague “my internet is slow” tickets become a ranked root cause, the log lines that prove it, and the next action.
Result: 7 analysis passes · 3 vector memories
View case studyService · Signals & sensors
We build AI that reads machine data the way senior engineers do: fault diagnosis and anomaly detection over device logs, network telemetry, sensor streams and power-system waveforms, on a server or on the device itself.
Selected work
Diagnostics and predictive models our founding team has shipped or published, from WiFi fleets to power transmission lines. Open any project for the full workflow and results.
Network & device telemetry2 projects
Telecom · home networking
Vague “my internet is slow” tickets become a ranked root cause, the log lines that prove it, and the next action.
Result: 7 analysis passes · 3 vector memories
View case studySmart home · telecom
Routers detect human activity without cameras, fully on the device. A privacy-friendly feature with no cloud cost.
Result: 93.5% test accuracy · 7 activities
View case studyPower & infrastructure3 projects
Power systems · transmission
Breakers reclose only when it’s safe, keeping lines in service and avoiding damage. Published in IEEE Transactions on Power Delivery.
Result: 2,442 simulated fault cases
View case studyPower systems · transmission
A fault-detection model that keeps working on power lines it was never trained on. Published in IEEE Access.
Result: 90.5% on unseen lines (GRU: 80.7%)
View case studyUrban infrastructure · climate risk
A fast surrogate for flood simulation, so planners can test far more rainfall scenarios.
Result: CatBoost surrogate · validated on held-out scenarios
View case studyClient names are withheld unless public. References are available on request.
All case studiesWhat this covers
Systems that ingest raw device and network logs and produce ranked, human-readable diagnoses with the evidence lines that support them.
Baselines per device model and firmware, drift and outage detection across fleets of thousands to millions of endpoints.
RAG over device manuals, past incidents, and vendor documentation so support engineers resolve faults without escalating.
The unglamorous plumbing done right: parsing zoo-like log formats, sessionizing, and feature stores that keep models honest.
Network equipment vendors, ISPs and carriers, smart-home and IoT platforms, and operations teams drowning in device logs. Our founding team has shipped AI diagnostics in telecom-grade WiFi environments, so we know what carrier reliability expectations feel like.
Common questions
No, it is the normal starting condition. Parsing and normalizing heterogeneous log formats is treated as first-class engineering work in the engagement, not an assumption we make and skip.
Not necessarily. Diagnosis pipelines can run on-premise or in your VPC, and we design data flows with carrier-grade privacy expectations in mind.
Yes, within honest limits. We typically split the system into a fast streaming layer for detection and a slower reasoning layer for diagnosis, so alerts are timely and explanations are thorough.
Start a project
Send a few lines about your data and the outcome you want. We reply within 1 business day with a feasibility read and a proposed scoping plan.