Full-StackData AnalysisData Engineering★ FeaturedContains private data

Tractafric Financial Process Automation

Led end-to-end development of a two-part solution automating Tractafric Equipment's manual financial processes, during a 6-month internship as Data Analyst & Full Stack Developer. First, a PFE proof-of-concept validated methodological approaches for performance/profitability analysis and predictive customer segmentation. This was followed by a production microservice web app (Django, PostgreSQL, ReactJS) that standardizes data and automates support computing and claims generation, transforming a multi-day manual workflow into a real-time process, alongside business intelligence reporting on profitability, support programs, and customer segmentation.

Technologies

DjangoReactJSPostgreSQL

Gallery

More like this

★

Sales To User Data Q&A Assistant Agent

Local-LLM chat assistant that turns a plain-language sales question into an auditable SQL query against 3.1M real Caterpillar-dealer records with guardrails built around five verified data traps that silently return a wrong number if ignored.

PythonFastAPIDuckDBOllama (Local LLM)SQLsqlglotJavaScript,SSE

Private repoView details

★

Portfolio website with CMS

A full-stack portfolio site built using: Next.js, an admin CMS to manage every section myself, and real English/French/Arabic support, right-to-left layout included.

Next.jsSupabasePostgreSQLTypeScriptTailwind CSS,Prismanext-intl

Private repoLive

★

At-Risk Churn Customer Account Detector

Churn-risk ranking engine dealer's aftermarket business, isolating a credible 230-account cohort from a 3.1M-row export corrupted by a row-cap bug that silently dropped 63% of customers, and beating recency/frequency baselines with a bootstrap-validated 0.51 PR-AUC.

Pythonscikit-learnDuckDBMLflowEvidentlypandasStreamlit

Private repoView details