About me

News

Nov 2025: Selected as Fellow at NXTP Ventures.
Aug 2025: Teaching Fellow for AC215 (MLOps & LLMOps) at Harvard SEAS.
Apr 2025: Machine Learning Researcher at the Kempner Institute, Harvard — working on RL with Verifiable Rewards (RLVR).
Jun–Sep 2025: Fellow at AGI House, San Francisco Bay Area.
Sep 03, 2024: I started my master's in Computational Science and Engineering at Harvard University.
Nov 22, 2023: I accepted an offer to join deco.cx to work on ML-powered personalization.
Jun 20, 2023: I interned at Google for the second time, on the YouTube team.
Jan 09, 2023: I interned at Meta for the second time, on the App Ads Platform team.

Hey! I am a second-year Computational Science and Engineering master's student at Harvard University, working in Reinforcement Learning with Verifiable Rewards (RLVR), advised by Sham Kakade. I completed my undergraduate degree in Computer Engineering at the Federal University of Paraíba (UFPB) in Brazil.

At Harvard, I’m a Teaching Fellow for AC215 (MLOps & LLMOps) and a Machine Learning Researcher at the Kempner Institute, focusing on Reinforcement Learning with Verifiable Rewards (RLVR). Previously, I interned twice at Google (YouTube and Search) and twice at Meta (AI for AR and Ads).

During my undergraduate degree, I was a member of the Artificial Intelligence Applications Laboratory (ARIA Lab) at UFPB. There, I contributed to projects and research (published two papers) and wrote my thesis on LLM-powered e‑commerce experiences.

In addition to my academic pursuits, I co-founded TAIL (Technology and Artificial Intelligence League), a non-profit organization oriented by the ARIA Lab professors. The academic league aims to teach new students, produce content, contribute to the AI and Data Analysis field, and help the members achieve their goals (internships, graduation programs, and full-time jobs). It currently has more than 65 active members throughout Brazil and other countries.

On the industry side, I have interned twice at Google (YouTube and Search), twice at Meta (AI for AR and Ads), and I have experience in startups (Praso, Atoptima, and deco.cx - where I worked for almost 2 years). I am also a Fundação Estudar's Tech Fellow (One of 1.5% approved over thousands of applications in 2021).

I like running, playing football (the one and only ⚽), surfing, and reading in my spare time.

You can get in touch with me at itamardprf at gmail.com, visit my GitHub, or connect with me in Linkedin. You can learn more about my career in both Linkedin and my resume (which may be outdated).

Extra Curricular Experience

Tech Fellow - Fundação Estudar

As a Tech Fellow, I contributed to the Fundação Estudar's community by helping students with their career paths, giving lectures and workshops, and helping the organization with their projects. I have also conducted tech talks with technology professionals to show the students the different career paths in the field and show discussions around specific topics: Amazon Student Recruiter, VTEX CTO, Itau technology superintendent, Tips for tech interviews preparation, etc.

TAIL - Co-Founder and President

TAIL (Technology and Artificial Intelligence League) is a non-profit Academic League held at UFPB that aims to support the students knowledge sharing and to teach new students from all courses (not only the ones in the Computer Science, Engineering and Data Science department). The organization is advised by the Phds and professors at UFPB: Telmo Filho, Thais Gaudencio and Yuri Malheiros (responsibles for the ARIA laboratory). The organization has more than 60 active members throughout Brazil and other countries.

Logic Circuits I - Boolean Algebra - assistant instructor

A freshman class at UFPB that has the highest failling rates. Throughout the course I gave lectures and helped students with exercises and projects. Our work was granted a teaching initiation award by the University after the student pass rate increased from 60% to 90%.

Projects Highlights

GANdido Portinari

Machine Learning class project at UFPB - 2020

Me and my colleagues used a CycleGAN to generate images that are similar to the paintings of Candido Portinari, a famous brazilian painter.

Capy Running Coach

Capy Running Coach

AC215 Best of Show — 2024

An AI-powered running coach on WhatsApp, combining RAG, fine-tuned LLMs, vector search, and a GKE deployment. Read the write-up for system design and results.

Bass tab generator

AI class project at UFPB - 2020

A project using LSTM networks to generate new bass tabs based on the Rage Against the Machine, Red Hot Chilli Peppers and Primus repertory.

Binary Bot

AI class project at UFPB - 2020

Consists in a Bot that trade options contracts based on previous data using a Time series approach in EUR/USD.

Medway Student Allocation

Project of Operations Research Class - 2020

Medway is a middle-sized (300 students) course at João Pessoa - PB. The problem here consists in allocating each student to one of the 15 advisors based on the time needed by the student and the estimated compatibility between each pair to let things balanced and not overload any advisor. The project was done using IBM CPLEX and some preprocessig libraries.

Helper Brasil Conta Comigo

The Brasil Conta Comigo is a program from the Brazilian Government to allocate students in med school final years into the public health sector to help with the COVID-19 Crisis. This project intuits to make it easier for these students to find their preferred places, filtering the data with the distance and the sector of work.

Learn more about me

You can find me at any of the links below

Download my resume • Connect on LinkedIn