Bruno Ismaël Razafindratsimba

Data · Analytics · Engineering

Chargement en cours…

DATA ANALYST · DATA ENGINEER · DATA SCIENCE

Making data truly useful.

From raw streams to actionable decisions: I integrate, clean, and model data to build clear, reliable tools that empower business growth.

Explore case studies Get in touch

DATA LINEAGE / 01DE LA SOURCE À L'USAGE

ANTANANARIVO, MADAGASCARINGEST → TRANSFORM → UNDERSTAND → DECIDESCROLL

01 BR / PORTFOLIO DATA

AREAS OF EXPERTISE

From initial ingestion
to actionable metrics.

A business-driven approach bridging data analytics, robust pipelines, and operational intelligence.

Analytics & Business Intelligence

Clear indicators, consistent reporting, and decision-driven interactive dashboards.

Power BIDAXExcelPower QuerySQL

Data Pipelines & Integration

Reliable automated ETL pipelines with high standards for data quality and integrity.

PythonPolarsPostgreSQLNiFiDocker

Applied Data Science

Statistical models and machine learning evaluated against concrete business challenges.

PythonKNNscikit-learnTensorFlow / Keras

Automation & Document AI

Targeted workflows and document RAG assistants grounded in verifiable sources.

n8nRAGTinyLlamaChromaDB

02 BR / PORTFOLIO DATA

SELECTED PROJECTS

Context first.
Methodology next.

Deliverables from real missions and documented projects, focusing on measured outcomes and clean architecture.

01Data Engineering

CASE / 01

In progress · Pipeline

ATG / TankMaster, from Stream to Table

Context: Data from ATG and TankMaster needed to feed an analytical core for operations and fuel volume monitoring.
Approach: Integration orchestration with Apache NiFi, Python & Polars transformations, persistent storage in PostgreSQL.
Deliverable / Status: Containerized Docker pipeline deployed. Automated volume checks and reconciliation tables created.

Apache NiFi
Python
Polars
PostgreSQL
Docker

02Data Engineering

CASE / 02

Completed · Production

ATG & JDE Automated Reconciliation

Context: Cross-referencing telemetry data from ATG with enterprise records in JDE to highlight operational discrepancies.
Approach: Automated reconciliation workflows with fuzzy matching and variance threshold alerting.
Deliverable / Status: Automated discrepancy reporting reducing daily audit time across storage tanks.

ATG
JDE
PostgreSQL
Python
Docker

03Analytics & BI

CASE / 03

Completed · Internship

ERP Data to Executive Dashboards (MGBI)

Context: During Data Analyst internship at Madagascar Business Intelligence: extracting from Odoo & EBP ERP systems.
Approach: ETL with advanced Excel and Power Query linked directly to PostgreSQL and SSMS databases.
Deliverable / Status: Delivered unified operational dashboards used weekly by finance and operations teams.

Odoo
EBP
Excel
Power Query
PostgreSQL
SSMS

04Analytics & BI

CASE / 04

Completed

Sales Intelligence with Power BI

Context: Transforming multi-channel sales figures into interactive performance metrics for leadership.
Approach: Data modeling in Power Query, star schema design, and advanced DAX measures for YoY comparisons.
Deliverable / Status: Comprehensive executive dashboard showing margins, store performance, and sales rep KPIs.

Power BI
Power Query
DAX
SQL

05Analytics & BI

CASE / 05

Completed

Travel Agency Booking Analysis

Context: Providing transparent visibility into booking flows, seasonality, and profitability for a travel agency.
Approach: Advanced Excel modeling, Pivot tables, Power Query automation, and Power Pivot dimensional models.
Deliverable / Status: Interactive multi-year booking and margin analyzer delivered to management.

Excel
Power Query
Power Pivot
TCD

06Data Science

CASE / 06

Completed · ML Project

Credit Card Anomaly & Fraud Detection

Context: Detecting fraudulent bank card transactions modeled as an imbalanced binary classification problem.
Approach: Exploration of KNN algorithm with NumPy, pandas, Matplotlib, TensorFlow / Keras, and a Gradio demo.
Deliverable / Status: Trained classification model exposed through an interactive real-time scoring web interface.

KNN
Python
NumPy
pandas
Matplotlib
Gradio

07AI & Automation

CASE / 07

RAG Prototype

Document Q&A Assistant with Citations

Context: Exploring an intelligent question-answering assistant grounded strictly in verified company documentation.
Approach: RAG architecture combining TinyLlama, ChromaDB vector storage, Hugging Face Transformers, and Trafilatura.
Deliverable / Status: Working local prototype providing accurate grounded answers with direct document citations.

TinyLlama
ChromaDB
Transformers
Trafilatura

08AI & Automation

CASE / 08

Completed · Workflow

Automated Email Validation & Dispatch

Context: Automating incoming email classification, approval circuits, and contextual response dispatch.
Approach: End-to-end event-driven automation workflow orchestrated in n8n with conditional filters.
Deliverable / Status: Operational routing pipeline handling alerts and reducing manual ticket assignment.

n8n
Automation
Email
Webhooks

09Data Science

CASE / 09

Completed · ML Project

Image Classification via CNN

Context: Building an image classification model accompanied by a live interactive web demonstration.
Approach: Convolutional Neural Network (CNN) trained with TensorFlow / Keras, evaluated with confusion matrices.
Deliverable / Status: End-to-end pipeline with preprocessing, augmentation, and Gradio testing UI.

CNN
TensorFlow
Keras
Gradio

10Data Science

CASE / 10

Completed · ML Project

Customer Persona & Loyalty Clustering

Context: Analyzing customer behavioral metrics to identify distinct clusters associated with brand retention.
Approach: Unsupervised K-means clustering with elbow method optimization and PCA dimensionality reduction.
Deliverable / Status: Identified 4 distinct consumer segments to guide targeted marketing campaigns.

K-means
Clustering
Python
scikit-learn

11Data Science

CASE / 11

Completed · Full-Stack ML

Healthcare Diagnostic Prototype

Context: Exploring risk classification applied to healthcare metrics with an accessible practitioner interface.
Approach: Neural network model served via lightweight Flask REST API and a responsive React frontend.
Deliverable / Status: Functional end-to-end demonstration prototype for clinical parameter exploration.

TensorFlow
Keras
Flask
React
Sass

12Data Science

CASE / 12

Completed · Full-Stack ML

Personalized Nutrition Recommender

Context: Recommending tailored dietary plans based on individual physical profiles and wellness goals.
Approach: Regression & constraint satisfaction algorithms exposed via Flask API and React frontend.
Deliverable / Status: Interactive nutrition generator providing macro-nutrient breakdowns and meal plans.

Python
Flask
React
Sass

03 BR / PORTFOLIO DATA

TOOLS & PRACTICES

The stack follows
the need.

Proficiency levels based on enterprise missions, client projects, and academic research.

Practiced in production & enterprise missions

PythonApache NiFiApache HopTalend DIPostgreSQLPower BI Desktop/ServicePower QueryDAXPower AutomateSFTPDockerSQLExcel avancén8n

Currently exploring & mastering

Talend ESBLangChainLangGraph

Documented in academic & full-stack projects

KNNCNNTensorFlow / KerasGradioChromaDBTransformersMongoDBGraphQLApache AirflowApache Kafka

JavaScript, React, Node.js, NestJS, and GraphQL practiced across production software development projects.

04 BR / PORTFOLIO DATA

PROFESSIONAL EXPERIENCE

Learning by
delivering.

Missions and internships, from enterprise data infrastructure to applied artificial intelligence.

3rd Place — Hackathon DevHub
Accès Banque · Team Zenith
June 2025

Participant — Hackathon DevFest 2025
Team API Leka
Dec 6–7, 2025

Participant — EMIHack 3.0
Internal EMIT Hackathon · Team API Leka
May 2025

Currently

Data & BI Intern

Logistique Pétrolière SA

Data Engineering and Business Intelligence internship in the petroleum logistics sector.

PythonApache NiFiPostgreSQLDocker

Apr — June 2026

Data Analyst Intern

MGBI · Madagascar Business Intelligence

Extracting from Odoo & EBP ERP systems, transformation and loading into reporting dashboards.

ExcelPower QueryPostgreSQLSSMS

Oct — Nov 2025

AI Engineer

GIRIS · AI Agency · Remote

Contributions to Computer Vision projects: model development, deployment, and optimization in applications.

PythonTensorFlowKerasGradio

June — Sept 2024

Software Engineering Intern

DSI · Ministry of Interior

Built a web application for traffic violation tracking and case management across Madagascar.

ReactNodeJSNestJSGraphQLMongoDB

June — Aug 2023

Web Development Intern

H. Création

Designed and built an online digital voting management web application.

HTMLCSSJavaScriptPHPJQueryMySQL

05 BR / PORTFOLIO DATA

BRUNO ISMAËL RAZAFINDRATSIMBA

Engineering in service of
the right questions.

Data Analyst and pipeline-oriented Data Engineer, pursuing a Master's Degree in Data Science & Artificial Intelligence at EMIT, University of Fianarantsoa.

EDUCATION & ACADEMIC BACKGROUND

2025 — 2026

Master II · Data Science & Artificial IntelligenceEMIT · University of Fianarantsoa

2024 — 2025

Master I · Data Science & Artificial IntelligenceEMIT · University of Fianarantsoa

2023 — 2024

Bachelor's Degree in Computer ScienceEMIT · University of Fianarantsoa

2019 — 2020

General Scientific Baccalaureate (Series C)Lycée Notre Dame

LANGUAGES

Malagasy Native

French Fluent / Professional

English Intermediate · technical documentation & specs

HAVE A DATA INQUIRY?

Let's discuss
your next pipeline.

A pipeline to stabilize, a dashboard to engineer, or a machine learning prototype to explore?

WRITErazafindratsimbabrunoismael@gmail.com CALL+261 32 95 729 93 / 38 94 055 22 CODE & PROJECTSgithub.com/IsmaelBruno

Antananarivo, Madagascar

Bruno Ismaël Razafindratsimba

Data · Analytics · Engineering

Chargement en cours…

DATA ANALYST · DATA ENGINEER · DATA SCIENCE

Making data truly useful.

From raw streams to actionable decisions: I integrate, clean, and model data to build clear, reliable tools that empower business growth.

Explore case studies Get in touch

DATA LINEAGE / 01DE LA SOURCE À L'USAGE

ANTANANARIVO, MADAGASCARINGEST → TRANSFORM → UNDERSTAND → DECIDESCROLL

01 BR / PORTFOLIO DATA

AREAS OF EXPERTISE

From initial ingestion
to actionable metrics.

A business-driven approach bridging data analytics, robust pipelines, and operational intelligence.

Analytics & Business Intelligence

Clear indicators, consistent reporting, and decision-driven interactive dashboards.

Power BIDAXExcelPower QuerySQL

Data Pipelines & Integration

Reliable automated ETL pipelines with high standards for data quality and integrity.

PythonPolarsPostgreSQLNiFiDocker

Applied Data Science

Statistical models and machine learning evaluated against concrete business challenges.

PythonKNNscikit-learnTensorFlow / Keras

Automation & Document AI

Targeted workflows and document RAG assistants grounded in verifiable sources.

n8nRAGTinyLlamaChromaDB

02 BR / PORTFOLIO DATA

SELECTED PROJECTS

Context first.
Methodology next.

Deliverables from real missions and documented projects, focusing on measured outcomes and clean architecture.

01Data Engineering

CASE / 01

In progress · Pipeline

ATG / TankMaster, from Stream to Table

Context: Data from ATG and TankMaster needed to feed an analytical core for operations and fuel volume monitoring.
Approach: Integration orchestration with Apache NiFi, Python & Polars transformations, persistent storage in PostgreSQL.
Deliverable / Status: Containerized Docker pipeline deployed. Automated volume checks and reconciliation tables created.

Apache NiFi
Python
Polars
PostgreSQL
Docker

02Data Engineering

CASE / 02

Completed · Production

ATG & JDE Automated Reconciliation

Context: Cross-referencing telemetry data from ATG with enterprise records in JDE to highlight operational discrepancies.
Approach: Automated reconciliation workflows with fuzzy matching and variance threshold alerting.
Deliverable / Status: Automated discrepancy reporting reducing daily audit time across storage tanks.

ATG
JDE
PostgreSQL
Python
Docker

03Analytics & BI

CASE / 03

Completed · Internship

ERP Data to Executive Dashboards (MGBI)

Context: During Data Analyst internship at Madagascar Business Intelligence: extracting from Odoo & EBP ERP systems.
Approach: ETL with advanced Excel and Power Query linked directly to PostgreSQL and SSMS databases.
Deliverable / Status: Delivered unified operational dashboards used weekly by finance and operations teams.

Odoo
EBP
Excel
Power Query
PostgreSQL
SSMS

04Analytics & BI

CASE / 04

Completed

Sales Intelligence with Power BI

Context: Transforming multi-channel sales figures into interactive performance metrics for leadership.
Approach: Data modeling in Power Query, star schema design, and advanced DAX measures for YoY comparisons.
Deliverable / Status: Comprehensive executive dashboard showing margins, store performance, and sales rep KPIs.

Power BI
Power Query
DAX
SQL

05Analytics & BI

CASE / 05

Completed

Travel Agency Booking Analysis

Context: Providing transparent visibility into booking flows, seasonality, and profitability for a travel agency.
Approach: Advanced Excel modeling, Pivot tables, Power Query automation, and Power Pivot dimensional models.
Deliverable / Status: Interactive multi-year booking and margin analyzer delivered to management.

Excel
Power Query
Power Pivot
TCD

06Data Science

CASE / 06

Completed · ML Project

Credit Card Anomaly & Fraud Detection

Context: Detecting fraudulent bank card transactions modeled as an imbalanced binary classification problem.
Approach: Exploration of KNN algorithm with NumPy, pandas, Matplotlib, TensorFlow / Keras, and a Gradio demo.
Deliverable / Status: Trained classification model exposed through an interactive real-time scoring web interface.

KNN
Python
NumPy
pandas
Matplotlib
Gradio

07AI & Automation

CASE / 07

RAG Prototype

Document Q&A Assistant with Citations

Context: Exploring an intelligent question-answering assistant grounded strictly in verified company documentation.
Approach: RAG architecture combining TinyLlama, ChromaDB vector storage, Hugging Face Transformers, and Trafilatura.
Deliverable / Status: Working local prototype providing accurate grounded answers with direct document citations.

TinyLlama
ChromaDB
Transformers
Trafilatura

08AI & Automation

CASE / 08

Completed · Workflow

Automated Email Validation & Dispatch

Context: Automating incoming email classification, approval circuits, and contextual response dispatch.
Approach: End-to-end event-driven automation workflow orchestrated in n8n with conditional filters.
Deliverable / Status: Operational routing pipeline handling alerts and reducing manual ticket assignment.

n8n
Automation
Email
Webhooks

09Data Science

CASE / 09

Completed · ML Project

Image Classification via CNN

Context: Building an image classification model accompanied by a live interactive web demonstration.
Approach: Convolutional Neural Network (CNN) trained with TensorFlow / Keras, evaluated with confusion matrices.
Deliverable / Status: End-to-end pipeline with preprocessing, augmentation, and Gradio testing UI.

CNN
TensorFlow
Keras
Gradio

10Data Science

CASE / 10

Completed · ML Project

Customer Persona & Loyalty Clustering

Context: Analyzing customer behavioral metrics to identify distinct clusters associated with brand retention.
Approach: Unsupervised K-means clustering with elbow method optimization and PCA dimensionality reduction.
Deliverable / Status: Identified 4 distinct consumer segments to guide targeted marketing campaigns.

K-means
Clustering
Python
scikit-learn

11Data Science

CASE / 11

Completed · Full-Stack ML

Healthcare Diagnostic Prototype

Context: Exploring risk classification applied to healthcare metrics with an accessible practitioner interface.
Approach: Neural network model served via lightweight Flask REST API and a responsive React frontend.
Deliverable / Status: Functional end-to-end demonstration prototype for clinical parameter exploration.

TensorFlow
Keras
Flask
React
Sass

12Data Science

CASE / 12

Completed · Full-Stack ML

Personalized Nutrition Recommender

Context: Recommending tailored dietary plans based on individual physical profiles and wellness goals.
Approach: Regression & constraint satisfaction algorithms exposed via Flask API and React frontend.
Deliverable / Status: Interactive nutrition generator providing macro-nutrient breakdowns and meal plans.

Python
Flask
React
Sass

03 BR / PORTFOLIO DATA

TOOLS & PRACTICES

The stack follows
the need.

Proficiency levels based on enterprise missions, client projects, and academic research.

Practiced in production & enterprise missions

PythonApache NiFiApache HopTalend DIPostgreSQLPower BI Desktop/ServicePower QueryDAXPower AutomateSFTPDockerSQLExcel avancén8n

Currently exploring & mastering

Talend ESBLangChainLangGraph

Documented in academic & full-stack projects

KNNCNNTensorFlow / KerasGradioChromaDBTransformersMongoDBGraphQLApache AirflowApache Kafka

JavaScript, React, Node.js, NestJS, and GraphQL practiced across production software development projects.

04 BR / PORTFOLIO DATA

PROFESSIONAL EXPERIENCE

Learning by
delivering.

Missions and internships, from enterprise data infrastructure to applied artificial intelligence.

3rd Place — Hackathon DevHub
Accès Banque · Team Zenith
June 2025

Participant — Hackathon DevFest 2025
Team API Leka
Dec 6–7, 2025

Participant — EMIHack 3.0
Internal EMIT Hackathon · Team API Leka
May 2025

Currently

Data & BI Intern

Logistique Pétrolière SA

Data Engineering and Business Intelligence internship in the petroleum logistics sector.

PythonApache NiFiPostgreSQLDocker

Apr — June 2026

Data Analyst Intern

MGBI · Madagascar Business Intelligence

Extracting from Odoo & EBP ERP systems, transformation and loading into reporting dashboards.

ExcelPower QueryPostgreSQLSSMS

Oct — Nov 2025

AI Engineer

GIRIS · AI Agency · Remote

Contributions to Computer Vision projects: model development, deployment, and optimization in applications.

PythonTensorFlowKerasGradio

June — Sept 2024

Software Engineering Intern

DSI · Ministry of Interior

Built a web application for traffic violation tracking and case management across Madagascar.

ReactNodeJSNestJSGraphQLMongoDB

June — Aug 2023

Web Development Intern

H. Création

Designed and built an online digital voting management web application.

HTMLCSSJavaScriptPHPJQueryMySQL

05 BR / PORTFOLIO DATA

BRUNO ISMAËL RAZAFINDRATSIMBA

Engineering in service of
the right questions.

Data Analyst and pipeline-oriented Data Engineer, pursuing a Master's Degree in Data Science & Artificial Intelligence at EMIT, University of Fianarantsoa.

EDUCATION & ACADEMIC BACKGROUND

2025 — 2026

Master II · Data Science & Artificial IntelligenceEMIT · University of Fianarantsoa

2024 — 2025

Master I · Data Science & Artificial IntelligenceEMIT · University of Fianarantsoa

2023 — 2024

Bachelor's Degree in Computer ScienceEMIT · University of Fianarantsoa

2019 — 2020

General Scientific Baccalaureate (Series C)Lycée Notre Dame

LANGUAGES

Malagasy Native

French Fluent / Professional

English Intermediate · technical documentation & specs

HAVE A DATA INQUIRY?

Let's discuss
your next pipeline.

A pipeline to stabilize, a dashboard to engineer, or a machine learning prototype to explore?

WRITErazafindratsimbabrunoismael@gmail.com CALL+261 32 95 729 93 / 38 94 055 22 CODE & PROJECTSgithub.com/IsmaelBruno

Antananarivo, Madagascar

Bruno Ismaël Razafindratsimba

Making data truly useful.

From initial ingestionto actionable metrics.

Analytics & Business Intelligence

Data Pipelines & Integration

Applied Data Science

Automation & Document AI

Context first.Methodology next.

ATG / TankMaster, from Stream to Table

ATG & JDE Automated Reconciliation

ERP Data to Executive Dashboards (MGBI)

Sales Intelligence with Power BI

Travel Agency Booking Analysis

Credit Card Anomaly & Fraud Detection

Document Q&A Assistant with Citations

Automated Email Validation & Dispatch

Image Classification via CNN

Customer Persona & Loyalty Clustering

Healthcare Diagnostic Prototype

Personalized Nutrition Recommender

The stack followsthe need.

Practiced in production & enterprise missions

Currently exploring & mastering

Documented in academic & full-stack projects

Learning bydelivering.

Data & BI Intern

Data Analyst Intern

AI Engineer

Software Engineering Intern

Web Development Intern

Engineering in service ofthe right questions.

EDUCATION & ACADEMIC BACKGROUND

LANGUAGES

Let's discussyour next pipeline.

ABI-MADAGASCARBruno Ismaël Razafindratsimba / Data Enthusiast

Bruno Ismaël Razafindratsimba

Making data truly useful.

From initial ingestionto actionable metrics.

Analytics & Business Intelligence

Data Pipelines & Integration

Applied Data Science

Automation & Document AI

Context first.Methodology next.

ATG / TankMaster, from Stream to Table

ATG & JDE Automated Reconciliation

ERP Data to Executive Dashboards (MGBI)

Sales Intelligence with Power BI

Travel Agency Booking Analysis

Credit Card Anomaly & Fraud Detection

Document Q&A Assistant with Citations

Automated Email Validation & Dispatch

Image Classification via CNN

Customer Persona & Loyalty Clustering

Healthcare Diagnostic Prototype

Personalized Nutrition Recommender

The stack followsthe need.

Practiced in production & enterprise missions

Currently exploring & mastering

Documented in academic & full-stack projects

Learning bydelivering.

Data & BI Intern

Data Analyst Intern

AI Engineer

Software Engineering Intern

Web Development Intern

Engineering in service ofthe right questions.

EDUCATION & ACADEMIC BACKGROUND

LANGUAGES

Let's discussyour next pipeline.

From initial ingestion
to actionable metrics.

Context first.
Methodology next.

The stack follows
the need.

Learning by
delivering.

Engineering in service of
the right questions.

Let's discuss
your next pipeline.

ABI-MADAGASCAR
Bruno Ismaël Razafindratsimba / Data Enthusiast

From initial ingestion
to actionable metrics.

Context first.
Methodology next.

The stack follows
the need.

Learning by
delivering.

Engineering in service of
the right questions.

Let's discuss
your next pipeline.