Looking for a job

David Macêdo

@davidmacdo

Message

AI Principal Scientist at Predactiv who improved multimodal recommendations by 60% with custom representation learning.

Brazil

Message

At Predactiv, I architect end-to-end AI systems and multimodal model lifecycles in PyTorch and Hugging Face. I developed custom representation learning models that combine image, text, and tabular data into user behavior embeddings, improving recommendation performance by 60%.

I also designed RAG and multi-agent systems with LangGraph, LangChain, and LangSmith to automate audience generation and campaign workflows. These systems accelerated enterprise time-to-market by 3x and audience creation by 75%.

At Samsung, I led the R&D team behind Galaxy AI voice and multimodal subsystems. I helped deliver production Edge AI systems, reduced on-device inference latency by 80% with TensorFlow Lite, and improved NLP and speech-processing performance by more than 20%.

At CIn/UFPE, I directed technical execution across more than 100 AI initiatives and authored around 50 peer-reviewed papers. I also co-founded and taught the graduate Deep Learning curriculum at Federal University of Pernambuco for 7 years, and conducted visiting research at MILA.

Experience

Work history, roles, and key accomplishments

Current

AI Principal Scientist

Current

Predactiv

Feb 2024 - Present (2 years 9 months)

Architected and deployed end-to-end AI systems in PyTorch and Hugging Face, reducing experimentation cycle times by 40% and boosting recommendation performance by 60%. Led technical adoption of GenAI and agentic workflows, accelerating enterprise time-to-market by 3x and driving a 20% revenue lift.

PyTorch Hugging Face fastAPI Django Self Supervised Learning Multi Modal Models RAG LangGraph Lang Chain Langsmith LoRAs VLLM Vector Databases

AI Specialist Research Engineer

Samsung

Jan 2022 - Jan 2024 (2 years)

Led a team of 10 engineers to deliver production AI systems for Edge AI platforms, achieving an 80% latency reduction with TensorFlow Lite and 10x data pipeline acceleration. Standardized MLOps workflows using Docker, Kubeflow, and GitHub Actions.

TensorFlow Keras Tensorflow Lite Tf.Data TensorFlow I O NLP Speech Processing Stts TTS Docker Kubeflow GitHub Actions MLOps

AI Research Scientist

Center of Informatics (CIn/UFPE)

Jan 2018 - Dec 2022 (4 years 11 months)

Directed technical execution across 100+ AI initiatives, architecting and benchmarking 200+ models in Python and C/C++. Engineered deep learning pipelines in PyTorch, TensorFlow, and Keras, and deployed models on embedded systems and cloud platforms.

Python C C++PyTorch TensorFlow Keras Computer Vision NLP Speech Processing BERT GPT AWS GCP Azure

AI Professor

Federal University of Pernambuco (CIn/UFPE)

Jan 2016 - Dec 2022 (6 years 11 months)

Co-founded and delivered the graduate Deep Learning curriculum, mentoring 200+ graduate researchers. Authored ~50 peer-reviewed papers in top AI venues, advancing methodologies in multimodal learning and robust representation.

Deep Learning Generative AI PyTorch TensorFlow Computer Vision NLP Speech Processing Research Mentorship

Education

Degrees, certifications, and relevant coursework

Federal University of Pernambuco

Doctor of Philosophy, Computer Science

Grade: 4.00/4.00

Activities and societies: First place in admission process. Google Research Award: 'Robust Deep Learning'. Google Cloud Platform (GCP).

Doctor of Philosophy in Computer Science with focus on AI and Deep Learning. Graduated with a 4.00/4.00 GPA, ranked first in the admission process, and received a Google Research Award.

Federal University of Pernambuco

Master of Science, Computer Science

Grade: 4.00/4.00

Activities and societies: First place in admission process. Microsoft Research Award: 'Deep Learning For Speaker Recognition'. Microsoft Azure.

Master of Science in Computer Science with focus on AI and Deep Learning. Graduated with a 4.00/4.00 GPA, ranked first in the admission process, and received a Microsoft Research Award.

Federal University of Pernambuco

Bachelor of Science, Electrical Engineering

Grade: 9.56/10.0

Activities and societies: First place in admission process. Highest Academic Distinction (Best Overall Student).

Bachelor of Science in Electrical Engineering. Graduated with a 9.56/10.0 GPA, ranked first in the admission process, and received the Highest Academic Distinction.