BNPL Risk Engine
ROC-AUC 0.85 · real-time BNPL scoring API: FastAPI + LightGBM, SHAP explainability, Docker, CI.
Open to freelance missions · remote
I build and harden production data pipelines for fintechs. Data Engineer at Rubyx, a Belgian fintech; open to scoped freelance missions alongside.
Freelance data engineering for fintechs and data teams, remote
dbt, Airflow on Kubernetes and BigQuery: from raw ingestion to tested, documented marts.
Data quality frameworks, parity testing, CI gates and SQL optimization to cut BigQuery costs.
ML, computer vision or LLM features served through FastAPI, from raw data to deployment.
Data Engineer @ Rubyx · Master Big Data Analytics
$ airflow dags trigger whoami_dag
14:22:01 task identity → running
→ Ibrahima Gabar Diop · Data Engineer: dbt · BigQuery · Airflow on GCP
14:22:02 task role → running
→ Data Engineer @ Rubyx, a Belgian fintech · hybrid
14:22:02 task stack → running
→ dbt · BigQuery · Airflow on Kubernetes · GCP
14:22:03 task impact → running
→ dbt model refactoring · BigQuery cost optimization · stateful dbt retries on K8s
✓ whoami_dag: 4/4 tasks succeeded · open to freelance missions█
From raw data to decision: the pipeline behind every project
Rubyx 🇧🇪 Belgian fintech Senegal · Hybrid
Lending-as-a-service for microfinance institutions in West Africa.
PestAI Senegal
Agricultural AI startup I co-founded; I owned the entire computer-vision side.
Saturne IA · ComeUp Remote
Orange Business Services Senegal Dakar, Senegal
B2B telecom subsidiary of Orange in Senegal.
UNCHK Senegal
Sonatel Academy (ODC) Dakar, Senegal
Université de Thiès Thiès, Senegal
Algebra, analysis, probability, statistics, algorithms and databases: the theoretical base for data engineering.
Data engineering first; AI where it adds value.
ROC-AUC 0.85 · real-time BNPL scoring API: FastAPI + LightGBM, SHAP explainability, Docker, CI.
17 dbt tests on live World Bank data · Airflow → DuckDB → dbt marts for the 8 WAEMU countries.
148K-row loan dataset · incremental models, SCD Type 2, data contracts, macros, slim CI.
dbt feature engineering + nightly batch scoring on Airflow, with a versioned model and a PSI drift gate.
dbt warehouse for digital lending: DPD, PAR30/60/90, IFRS9-style provisioning, origination cohorts.
dbt Cloud orchestration from Airflow: manifest parsing, quality gates, Slack alerting.
Tested production Airflow patterns: dynamic task mapping, spot-node pods, cost-guarding policies.
CSV → DuckDB → dbt → Streamlit, orchestrated with Airflow; one-command startup.
AI × Data
More on github.com/Gblack98
NVIDIA Deep Learning Institute · Jul 2025
Neo4j GraphAcademy · Jan 2025
Sonatel Academy · Jan 2023 – Nov 2023
Force-N Senegal · Aug 2023 – Jan 2024
Tell me about your data stack or project.
How it works