🐍
🧠
🔥
🧭
☁️
🐳
🧱
🔗
⚙️
📈
Amine Bousmah
Data & AI Engineer
I turn data into decisions. I shape messy data into simple answers, work closely with teams, and ship things that truly help people. 🤝🚀
PythonSQLAirflowSparkCloud (GCP / AWS / Azure)DatabricksdbtSnowflakeBigQueryTensorFlowPyTorchscikit-learnMLflowDockerGitCI/CDAPIJiraPower BITableauKafkaC#Dataiku DSS
Scroll to explore
Technical Skills
🛠️
Data Engineering & Analytics
Foundations for reliable data systems.
ETL/ELT with Python & SQL; orchestration basics (dbt/Airflow).
Star/snowflake modeling and essential warehousing patterns.
Data validation and tests with clear SLAs/expectations.
Query optimization fundamentals and cost-aware thinking.
Small streaming prototypes (Kafka) + batch/stream joins.
🤖
Machine Learning & Modeling
Pragmatic ML with strong evaluation discipline.
Solid baselines (linear, tree-based) before complex models.
Time-series forecasting (ARIMA/Prophet, boosting) when useful.
Feature pipelines with leakage-safe cross-validation.
Explainability (SHAP/feature importances) and readable model cards.
Experiment tracking (MLflow) and packaging models for APIs.
📊
BI & Data Visualization
Make results clear, trusted, and actionable.
Defined KPIs and a simple semantic/metric layer.
Interactive dashboards with filters and drill-through.
Data storytelling: annotations, small multiples, clear legends.
Row-level security basics and governance-ready layouts.
Scheduled refreshes, exports, and lightweight QA checks.
🧩
Application Design & API
Product-minded developer focused on clean, secure services.
DDD-lite: clear module boundaries and dependency rules.
REST APIs (FastAPI/Flask) with typed schemas and OpenAPI.
Auth (JWT/OAuth2), input validation, and robust error handling.
Background jobs (Celery/RQ), file ingestion, async I/O.
Frontend integration with React/Next and reusable UI patterns.
☁️
Cloud & DevOps
Ship small, observe, and iterate.
Containerized dev with Docker; reproducible environments.
CI/CD (GitHub Actions): tests, linting, type checks.
Deploy on Vercel/Cloud Run; env & secrets management.
Basic monitoring (logs/metrics/traces) and alerting.
Cost awareness and usage-based scaling (serverless first).
💹
Data Finance & Revenue Analytics
Applied analytics for markets, risk, and growth.
Credit risk scorecards & PD estimation; calibration & backtesting.
Market analytics: returns/volatility, simple VaR/ES & stress tests.
Fraud/AML: anomaly detection and KYC/KYB entity matching.
Portfolio analytics: mean-variance, factor tilts, Black-Litterman basics.
Documentation & governance aligned with IFRS/Model Risk standards.
Selected Projects
Vinted Extension — Smart auto-repost to boost visibility
Browser extension that automatically republishes listings to leverage Vinted’s algorithmic boost. Features safe scheduling, anti-duplicate logic and local anti-tracking to maximize views and click-throughs without manual effort.
Key results
95%
automation
85%
time Saved
30%
view Uplift
Technical implementation
- ▹Chrome Extension (content + background service worker)
- ▹Task scheduler, de-duplication & cooldown management
- ▹Local headers/cookies handling; optional Express helper
- ▹Image helpers (crop/compress) when needed
Tribara — Talent Matching Optimization
AI-powered recruitment optimization to automate candidate screening and ranking, integrated with ATS for seamless workflows. Delivered faster shortlists and more relevant matches for recruiters.
Key results
50%
screening Time Reduction
500
cv Volume
30%
relevance Gain
Technical implementation
- ▹Python ETL & ML pipeline (parsing + scoring)
- ▹NLP-based candidate ranking with continuous fine-tuning
- ▹ATS integration (webhooks/API) & scoring feedback loop
- ▹Dashboard & export for recruiter decision support
Face Recognition — Find all photos of a person
Application that lets you upload a few photos of yourself to automatically detect all occurrences within an event album (ideal for team building/seminars).
Key results
91.4%
detection A P
99.83%
lfw Accuracy
10000
index Size
Technical implementation
- ▹Face detection: RetinaFace/SCRFD (InsightFace)
- ▹Face embeddings: ArcFace (InsightFace, 512-d vectors)
- ▹Similarity search & scaling: FAISS (IVF/PQ or HNSW)
- ▹De-duplication & robustness: thresholding + DBSCAN; multi-reference averaging
11Field — Football analytics & scouting suite
End-to-end scouting toolkit: xG/xGA, role-based radars, league comparators, match reports and player similarity. Adds ML models for clustering and explainability to support recruitment decisions.
Key results
12%
leagues Covered
40%
modeled Features
60%
time To Insight
Technical implementation
- ▹Data ingestion from public football APIs (FBref/ESPN/ClubElo, etc.)
- ▹Interactive dashboards (Streamlit + Plotly)
- ▹PCA + KMeans for playing-style clusters
- ▹RandomForest + SHAP for explainable player ranking
Modern Data Capabilities
🔌
Ingestion & Connectivity
- REST/GraphQL, webhooks, SaaS & DB connectors
- Batch files (CSV/Parquet) + CDC/event streams
- Secrets, retries, backoff, idempotency
🧭
Workflow Orchestration
- Reproducible DAGs with clear SLAs
- Idempotent tasks, alerts, backfills
- Data-aware scheduling & dependency management
🗄️
Lakehouse Storage & Formats
- Object storage + warehouse interoperability
- Parquet/Delta/Iceberg, partitioning & compaction
- Schema evolution, time travel & ACID tables
🏗️
Modeling & ELT
- Layered models (staging → core → marts)
- Data contracts & tests (quality as code)
- SCD patterns, surrogate keys, audit columns
🧪
Data Quality & Observability
- Freshness, completeness, accuracy monitors
- Column-level lineage & impact analysis
- Anomaly detection with playbooks/runbooks
📊
BI & Semantic Layer
- Governed metrics/semantic layer for consistency
- Row-level security & policy-based access
- Drill-through dashboards, alerts & subscriptions
✨
Data Apps & UX
- API-first apps (Next.js/React) with great UX
- Accessible, fast, mobile-friendly interfaces
- Shareable exports & decision-ready views
⚡
Realtime & Streaming
- CDC & event-driven pipelines (micro-batch/stream)
- Live dashboards via WebSockets/SSE
- Materialized views & low-latency caches
🤖
ML & MLOps
- Feature pipelines with reproducible training
- Experiment tracking, registry & versioning
- Drift/fairness monitoring & A/B evaluations
🧠
LLM & RAG
- Embeddings & chunking with prompt versioning
- Hybrid retrieval + guardrails & citations
- Privacy-aware grounding on enterprise data
🧭
Vector Search
- ANN indexes (HNSW, IVF-PQ) at scale
- Hybrid keyword + vector retrieval
- Deduplication & clustering for discovery
🔒
Governance, Privacy & Security
- RBAC/ABAC, masking/tokenization of PII
- Catalog, lineage, ownership & audit logs
- Compliance by design (GDPR, ISO 27001)
💸
FinOps & Performance
- Cost tags/budgets & storage lifecycle
- Pruning, partition pushdown, caching
- Autoscaling, SLAs/SLOs with clear error budgets
🚀
Data CI/CD & DevEx
- Git-based reviews, tests & linters for data code
- Reproducible builds & artifact versioning
- Ephemeral preview envs & safe rollbacks
🔁
Interop & Internal APIs
- OpenAPI/JSON Schema contracts & governance
- Reverse ETL to operational tools
- Pagination, rate limits & idempotent writes
Let's work together 🚀
Are you looking for a Data & AI profile capable of combining technical expertise with a human touch, someone who turns data into meaningful stories, builds solutions that matter, and works hand in hand with teams to create impact?
Paris, France