About

Bernardo de Lemos 👨🏻‍💻

Principal Machine Learning Engineer at McKinsey & Company, based in Portugal. My expertise lies in designing and implementing end-to-end Machine Learning and Generative AI systems, with a focus on real-time predictions, features, and ML lifecycle management.

Current Role & Impact 🎯

At McKinsey & Company, I lead the development of ML tools and systems that drive growth and pricing strategies in the B2B sector.

• Led the GenAI initiative for B2B growth
• Led architecture and development of ML tools for growth and pricing strategies
• Enhanced code optimization through custom profiling tools
• Created data preparation and validation tools

Professional Experience 💼

McKinsey & Company (2022-Present)

Principal Machine Learning Engineer (Promoted from Senior mid 2024) | B2B Pricing & Growth

• Led the GenAI initiative for B2B growth:
  • Developed foundational framework for serving GenAI capabilities
  • Implemented GenAI approaches (workflows and agents) for business strategy
• Led architecture and development of ML tools for growth and pricing strategies
• Designed and optimized multi-stage CI/CD pipelines
  • Implemented caching strategies and parallel job execution in GitHub Actions and CircleCI
  • Created cost-effective CI/CD strategies through runner optimization and workflow refinement
  • Implemented matrix testing across multiple Python versions and dependencies
  • Automated ML model testing, validation, and deployment processes
• Enhanced code optimization through custom profiling tools
  • Custom time-based profiler for code optimization
  • Memory profiler integration for resource optimization
  • Benchmarking framework for ML pipeline performance tracking
• Created data preparation and validation tools:
  • Column mapping and type coercion
  • CLI and programmatic interfaces

Farfetch (2021-2022)

Machine Learning Engineer | Recommendations

• Architected real-time product recommendation system and migration from batch to stream processing.
• Developed ML models serving millions of users
  • Modular recommender models architecture
  • Zero downtime model upgrades
• Collaborated with cross-functional teams on recommendation strategies

QOMPLX (2019-2021)

Quant Analyst | AI Capabilities

• Designed No Code ML platform for model training and serving
• Developed quantitative finance ML pipelines
• Created ETL/ELT pipelines using Kafka, Spark, Avro, and Parquet
• Engineered real-time APIs in Scala and Akka

Banco de Portugal (2017-2018)

Data Scientist | Microdata Research Lab

• Developed multiple ML and data mining models for dataset cleansing, standardization and anomaly detection.
• Created a small compiler for analyzing the Stata programming language and its outputs.
• Implemented data obfuscation techniques to protect sensitive data.

JUMIA (2016)

Software Engineer | Business Intelligence

• Developed a data manipulation web application in Ruby on Rails to streamline data access and editing.
• Created SQL queries and conducted business intelligence analysis to provide valuable insights for decision-making.

---

Technical Expertise 🛠️

Programming Languages

• Advanced: Python
• Intermediate: Rust, Go
• Basic: Scala, Prolog

Machine Learning & Data Science

• Frameworks: PyTorch, TensorFlow, Keras, scikit-learn, LangChain, llama-index
• MLOps: MLflow, Transformers
• Analysis: XGBoost, Pandas, Polars, Numpy
• Vector Stores: PGVector, Redis

Backend Development & Asynchronous Programming

• APIs: FastAPI and asyncio (Python), Actix-web and tokio (Rust)

Infrastructure & Tools

• Cloud: Databricks, AWS
• CI/CD: Docker, CircleCI, GitHub Actions
• Data Processing: Kafka, Spark, Airflow
• Databases: MongoDB, PostgreSQL, Redis

Specialized Knowledge 🧠

Machine Learning Systems

• Real-time prediction serving and feature engineering
• MLOps and model lifecycle management
• Inference optimization
• Performance profiling and optimization
• Data validation and pipeline automation

GenAI Applications

• Agent and Workflow development
• Prompt engineering and optimization
• RAG systems development
• B2B applications of generative AI

Backend Development

• API design and implementation with FastAPI (Python) and Actix-web (Rust)
• Scalable, and performant systems for ML serving and data processing

Data Engineering

• Real-time streaming architectures
• Batch and streaming data processing
• Data validation and quality assurance
• ETL/ELT pipeline design and implementation

Languages 🗣️

• Native: Portuguese
• Fluent: English
• Conversational: Spanish