Lucas Tramonte

Data Scientist | Machine Learning | ex-Amazon

Resume

About Me


Machine Learning Engineer with hands-on experience building and deploying ML pipelines, automating workflows, and delivering data-driven analysis with cross-industry adaptability to drive measurable business impact.

I hold a double degree from CentraleSupélec (France) and Unicamp (Brazil), where I built my background in AI, optimization, and Industrial Engineering. Currently, I’m working on building predictive solutions for the Revenue Squad at Aprix.

My experience spans diverse industrial contexts, from optimizing labor flexibility at Amazon in Luxembourg to automating hardware fault detection for Motorola at the Eldorado Institute. I also have a background in high-volume data integration and simulation, having worked with FlexSim to automate data imports and evaluate predictive frameworks to accelerate team adoption.

I am fluent in English and French and work primarily with Python, SQL, and AWS, focusing on delivering clear and actionable insights to stakeholders.

Projects


OpenFinanceAI

OpenFinanceAI preview

SARS-CoV-2 Maternal Impact

Intent Detection for Travel Chatbots

Intent Detection for Travel Chatbots

Radiological-pollution-monitoring-anomaly-detection

MDVRP-Magazine-Luiza

Smart Chair Posture Analysis

Smart Chair Posture Analysis

Welding-Quality-Prediction

Optimization-AirFrance

Prediction of the SWOT analysis

Swot Matrix

Experience


Aprix logo

Aprix

Associate Machine Learning Engineer - -

Full-time · Brazil (Remote)

  • Eliminated a major revenue overestimation blind spot by deploying an end-to-end ML pipeline (Logistic Regression + Pipedrive integration) that replaced non-audited sales assumptions with data-driven win probabilities for active deals, enabling accurate financial projections with a 8.13% error margin in Q1 2026 via automated CI/CD workflows.
  • Automated the end-to-end Identity and Access Management (IAM) lifecycle for corporate platforms and production PostgreSQL flexible servers, recovering approximately 8.4 hours of manual effort per week across a monthly volume of ~292 audit-tracked tickets, by designing a modular Python serverless solution on Azure Functions.

Working on the Governance and AI squad, building end-to-end AI and predictive solutions for internal teams.

Machine Learning MLOps Logistic Regression Pipedrive CI/CD
Aprix logo

Aprix

AI Trainee - -

Apprenticeship · Brazil (Remote)

  • Reduced incident triage latency by 5–10 minutes per ticket for the Customer Support team, protecting SLA commitments with enterprise clients such as Ipiranga and Seara, by automating the full routing pipeline in n8n to ingest OTRS tickets, extract client context, trigger real-time Discord alerts, and auto-generate Kanban cards.
  • Saved 3 hours/week in manual audits and prevented escalation delays by building an automated governance bot (n8n) that sweeps customer tickets, applying a multi-level prioritization matrix (severity, SLA status, and age) to instantly flag requests unanswered for 7+ days directly to account managers on Discord.
  • Improved planning under uncertainty across business units by replacing subjective deadline-risk discussions with probabilistic completion scenarios, by developing and running a Monte Carlo forecasting engine with automated data ingestion and standardized management reporting.

Aprix is a SaaS company providing AI-driven pricing solutions for major industries in Latin America. Our technology helps large-scale businesses optimize pricing strategies and operational efficiency through scalable data products.

Azure DevOps Services n8n Discord Webhooks Monte Carlo OTRS
Amazon logo

Amazon

Business Intelligence Engineer Intern - -

Internship · Luxembourg (On-site)

  • Led research and proof-of-concept in Amazon EU Supply Chain (middle-mile) to improve short-term labor planning guardrails for Sort Centers over 24-hour and 48-hour horizons.
  • Designed and validated predictive approaches under temporal dependencies, censored outcomes, class imbalance, and site-specific behavior using SQL-based data extraction and Python feature engineering.
  • Built and evaluated ML pipelines in AWS SageMaker Studio against operational heuristics and observed a 27.4% uplift over baseline for same-day forecasting across 10 UK Sort Centers.
  • Partnered with BI, Science, and Operations stakeholders through recurring technical and business reviews to translate model behavior into production-readiness trade-offs.
Python SQL AWS SageMaker Forecasting Supply Chain Analytics
CNPq logo

CNPq - Conselho Nacional de Desenvolvimento Científico e Tecnológico

Machine Learning Researcher - -

Part-time · Limeira, São Paulo, Brazil (On-site)

  • Awarded a research scholarship from Brazil's National Council for Scientific and Technological Development (CNPq), supported by BIOS (Brazilian Institute of Data Science).
  • Conducted research on interpreting machine learning predictions for diagnosing SARS-CoV-2 from symptoms, using SHAP and LIME methods.
  • Link to the project
Python Interpretability SHAP LIME
FlexSim logo

FlexSim Brasil

Data Scientist Intern - -

Internship · Campinas, São Paulo, Brazil (Hybrid)

  • Eliminated software crashes during large-dataset imports by engineering token-based batch processing of 50,000 rows per chunk, integrating SQL connections with C++ simulation logic.
  • Conducted an analysis of no-code machine learning frameworks, identified RapidMiner as the best viable option due to usability and feature set, and upskilled the engineering team with a case study on steel-industry energy consumption prediction using Random Forest classifiers.

FlexSim Brasil is a highly specialized tech consultancy leveraging advanced Digital Twin and 3D simulation platforms to optimize complex, large-scale industrial and logistics operations.

SQL Machine Learning C++ RapidMiner
Eldorado logo

Instituto de Pesquisas Eldorado

Data Scientist Intern - -

Internship · Campinas, São Paulo, Brazil (Remote)

  • Eliminated manual triage bottlenecks and improved fault-detection precision to 73% by replacing an unmaintainable, rule-based system with an end-to-end Machine Learning pipeline for battery telemetry anomaly detection.
  • Expanded domain experts’ diagnostic criteria with data-backed insights, leveraging Explainable AI (SHAP values) to translate complex, hidden telemetry patterns into stable, production-ready predictive signals.

Instituto de Pesquisas Eldorado is one of Latin America's premier R&D institutes (1,600+ engineers), partnering with global tech leaders such as Motorola to develop software, AI, and Industry 4.0 solutions.

Python Data Science SHAP

Education


Ecole CentraleSupélec, Université Paris-Saclay

Master of Engineering - MEng, Artificial Intelligence - -

  • CentraleSupelec (formerly Ecole Centrale Paris) is ranked 2nd in France's Engineering Schools (QS Ranking 2020) and is part of Universite Paris-Saclay, ranked 13th globally by the Shanghai Ranking (2025).
  • Granted BRAFITEC's Excellence Scholarship from CAPES, completed a Research Minor, and pursued a Double Degree with final-year specialization in Artificial Intelligence.
  • Relevant coursework: Artificial Intelligence, Machine Learning, Optimization, Machine Learning System Design, Advanced Programming and Software Development Tools, and Advanced Natural Language Processing with Deep Learning.

State University of Campinas - UNICAMP

Bachelor's degree, Industrial Engineering - -

  • UNICAMP is one of Brazil's leading public research universities and a major center for science and innovation. In the QS Latin America and Caribbean 2026 ranking, UNICAMP is 3rd in Latin America.
  • Activities and societies: Physics Instructor at Cursinho Pre-Vestibular Colmeia (6h/week) and Teaching Assistant (300+ hours in Calculus I, Linear Algebra, Analytic Geometry, and General Physics I).
  • Relevant coursework: Algorithms and Computer Programming, Statistics and Probability, Numerical Methods, Operations Research I and II, Multicriteria Decision Support Methods, Machine Learning, Systems Simulation, Economic Engineering, Quality Engineering, and Project Management.

Contact