Machaka Ground Station

Projects

Three missions; each includes focused Operations with telemetry.

Mission I — Airbus

Data Engineer • Jan 2023 – Jun 2024 • Toulouse

Minimise deployment time for analytics and move high-throughput data reliably for the OPTIMATE autonomous-taxiing programme.

A1OPTIMATE Streams

Feed autonomous taxiing with reliable, low-latency streams under high load.

  • Implemented C++ stream components sustaining ~2 GB/s under test.
  • Designed predictable backpressure + bounded queues.
  • Added scriptable replay paths; wrote integration notes for downstream teams.
Throughput: ~2 GB/sStability: clean failure modesRepro.: deterministic replays
Before

Ad-hoc ingestion, uncertain behavior under load

After

Single, documented ingestion path with known limits

A2Dashboard Deployment Acceleration

Reduce time-to-deploy dashboards by eliminating manual ETL steps.

  • Built Python ETL automation and standardized data contracts.
  • Packaged deployment scripts for consistent releases across teams.
Lead-time: −80%Consistency: fewer manual steps
Before

Snowflake releases, regional variation

After

Reproducible, scripted deployments

A3HR Data Unification

Improve HR data accuracy across 5 regions for downstream dashboards.

  • Integrated datasets; reconciled schemas; added idempotent transforms.
  • Surfaced diffs for quick fixes; documented schema versions and ownership.
Coverage: 5 regionsQuality: fewer reconciliation issues
Before

Fragmented sources, inconsistent KPIs

After

Unified feeds with clearer ownership

Mission II — Green Praxis

Cloud & Data Engineer • Jan 2025 – Present • Aix-en-Provence

Build reliable data infrastructure and internal tooling for environmental & geospatial products.

G1Geo Pipelines & Dynamic Tiles

Automate multi-band satellite imagery ingest and serve dynamic, reliable map tiles.

  • Designed Airflow DAGs for ingest/transform; versioned data contracts.
  • Shifted serving to dynamic Google Earth Engine backend; added guardrail checks.
Iteration: ships fasterRepro.: repeatable runsUX: reduced lag
Before

Manual/semi-manual tiling, drift between runs

After

Airflow-managed processing + dynamic serving

G2Observability Rollout

Provide end-to-end visibility across DAGs, APIs, and the Kubernetes cluster.

  • Deployed Prometheus + Grafana; standardized labels/owners; pruned noisy alerts.
  • Integrated Airflow task metrics and API latency/error budgets into a single view.
Alert cov.: ≈95%Awareness: faster MTTRClarity: owners + runbooks
Before

Fragmented monitoring, key failure modes missed

After

Unified dashboards; actionable alerts

G3Internal Tooling & Platform Hygiene

Reduce developer toil; make pipelines easier to evolve.

  • Templated DAG patterns; small CLI helpers for common ops.
  • Tightened Terraform modules for repeatable infra.
Toil: reducedSpeed: smoother onboarding
Before

Ad-hoc scripts

After

Reusable templates & helpers

Mission III — Murex Systems

Project Manager • May 2021 – Jan 2022 • Beirut

Deliver a log-analysis demonstrator for a high-frequency trading platform, shipping an MVP on time with a small team.

M1Log-Analysis PoC (HFT)

Detect unseen error patterns quickly to accelerate root-cause analysis.

  • Led a 3-developer Scrum team; Python pipeline to parse/cluster anomalies.
  • Iterated with users for actionable output.
Discovery: >50 patternsSpeed: faster troubleshootingDelivery: on time
Before

Long hunts over raw logs

After

Ranked signals with clear context

M2MVP Delivery in K8s

Package and deploy the MVP in a modern, reproducible environment.

  • Containerised services (Docker); orchestrated with Kubernetes; CI on Jenkins.
  • Wrote a pragmatic runbook.
Repro.: one-command spinsVelocity: faster iterationStability: fewer env issues
Before

Snowflake environments

After

Standardised containers + CI