Luke Howlett Python · Data · Systems Engineer

From complex signals to reliable systems.

Python engineer with professional experience across GNSS/PNT, large technical datasets, production observability and cloud infrastructure. I turn specialist analysis into tools teams can operate and trust.

Professional experience

GNSS research translated into operational engineering.

Techlett Consulting / DDK Positioning

GNSS / Data Engineering Contractor

2025–2026

Delivered Python analysis, infrastructure and operational monitoring for GNSS correction and PPP/SSR-related positioning systems.

  • Built and integrated Grafana, CloudWatch and PagerDuty monitoring for production positioning infrastructure.
  • Implemented alerting around stream disconnects, timing irregularities, network drops, poor performance, low satellite counts and database heartbeat failures.
  • Analysed 15.6M+ GNSS correction observations to investigate quality, behaviour and operational failure modes.
  • Contributed Python, Terraform and AWS delivery across Lambda, S3, SNS and RDS/PostgreSQL services.
PythonAWSPostgreSQLGrafanaPagerDutyTerraformPPP / SSR

CHC Tech · Norwich

GNSS Research & Data Analyst

2020–2024

Applied mathematical and data-analysis methods to high-precision positioning research, receiver evaluation and GNSS correction services.

  • Developed Python tooling for GNSS/PNT research, positioning performance analysis and repeatable engineering evaluation.
  • Worked with RTCM and RINEX data across receiver testing, PPP/SSR concepts and correction-service analysis.
  • Evaluated algorithms and large scientific datasets, translating research findings into clear technical evidence.
PythonNumPySciPypandasRTCMRINEXGNSS / PNT
GNSS Observatory dashboard showing global reference-station health Live standalone project

Featured project

GNSS Observatory

Interactive observability platform for monitoring global GNSS reference-station health.

A reproducible Python analytics pipeline and interactive dashboard that turns public IGS/CDDIS station products into explorable network health, anomaly and observation-readiness evidence.

Technologies
PythonGeoRustPRXReactTypeScriptStatic JSONDocker
Engineering evidence
  • Data engineering
  • GNSS processing
  • Observability
  • Geospatial visualisation
  • Static deployment
What you're seeing

A real, standalone engineering project built from public GNSS products and deterministic synthetic scaling.

Typical tools
PythonGNSSStatic analyticsInteractive dashboard
What this proves

I can take a specialist data problem from source processing through analytics, visualisation and production deployment.

02 · Challenge

Find degraded coverage quickly.

Large reference networks generate more observation-quality data than an engineer can review station by station. The interface prioritises emerging issues and gives each score an evidence trail.

03 · Pipeline

Python analytics to static JSON.

A Python workflow parses public products, computes station-level health measures and exports deterministic artefacts for a responsive React and TypeScript interface.

04 · Delivery

Backend-free at runtime.

The deployed dashboard consumes precomputed JSON, keeping the public surface simple while still demonstrating data engineering, geospatial visualisation and production deployment.

Technology

A full analytical product, not a static mock-up.

PythonGeoRust / PRXReactTypeScript MapLibre / deck.glPlotlyStatic JSONDocker

Data pipeline

From public observations to an operational interface.

  1. IGS / CDDISPublic station data
  2. GeoRust / PRXRINEX processing
  3. Python analyticsMetrics and health scoring
  4. Static JSONReproducible export
  5. React dashboardInteractive observability

Additional projects

More evidence across analysis, infrastructure and product engineering.

private UIjob queueisolated workspaceOpenClaw gateway
Problem
Prompt-driven file analysis needs stronger boundaries than a general chat interface.
What I built
A private job-based workbench with uploads, persistent execution, isolated per-job workspaces and downloadable outputs.
Engineering decisions
The web and worker services never receive the model API key; the authenticated gateway owns model access.
Proof
FastAPI, OpenClaw, NVIDIA NIM, Docker Compose, non-root services and automated checks.
Friend Hub isolated public demo showing real-time chat, a generated image and online presence
Problem
Useful group decisions, plans and media disappear into unstructured chat history.
What I built
A self-hosted social workspace combining room-scoped chat, events, polls, reminders, notes, search, media and push notifications.
Engineering decisions
HttpOnly sessions, authenticated media, isolated demo guests, WebSockets and deterministic archive imports.
Proof
React PWA, FastAPI, PostgreSQL, Docker, Caddy and a public synthetic-data demo.
Problem
Correction streams can drift, disagree or hide subtle biases that affect downstream positioning.
What I built
Python analysis tooling for repeatable comparison of correction behaviour, bias patterns and quality indicators.
Stack
Python, pandas, statistical plotting and RTCM/SSR data workflows.
Engineering value
Turns visual suspicion into measurable evidence for technical review.

Capabilities

Click a lane to connect skills with project evidence.

GNSS correction analysis quality-control pipelines Python data processing AWS operational systems Terraform infrastructure as code Grafana, CloudWatch and PagerDuty Docker and Nginx deployment React data interfaces technical analysis and communication

Education & details

A mathematical foundation for evidence-led engineering.

Based in

Chelmsford, UK

Available for UK remote and hybrid engineering work.

Professional status

Independent contractor

Techlett Consulting · UK registered sole trader.

Contact

Let’s talk engineering.

GNSS, Python, data systems, observability and infrastructure.