
- Joined
- Oct 2026
- How they joined
- Directly
- Before Cognition
- Sequoia Capital๐
- Status
- Current
- Joined Cognition in Oct 2026 on the 'Research Team + Office of the President' per his LinkedIn experience entry; Cosign lists him as Researcher at Cognition from Sep 2026, so the role reads as research plus chief-of-staff-style work rather than a pure Devin/Windsurf product-engineering seat.Oct 2026
- Shared Cognition's announcement of raising over $2B at a $48B valuation with the comment "do it all" on 2026-09-08, around the time he joined.Sep 8, 2026
- Sequoia Capital scout since Jan 2026; LinkedIn lists Sable, Sail Research, Universal Constructor, Intelligence.ai, MIRA, Tzafon, CompLabs, Pantheon and Creatorwood among companies.Jan 2026
- Member of Technical Staff at Thinking Machines Lab Dec 2025-Oct 2026; his description: "Trying to connect research <> product, promo in 3mo; Cooking on evals, RL, and a very, very cool product."Dec 2025
- Summer 2025 Member of Technical Staff at Google DeepMind on the Gemini Core Model Team, Frontier Evals.May 2025
- Did engineering at a stealth startup Jun-Dec 2024 building around AI agents, digital twins and cyberattacks; he states the company was acquired by depthfirst.com (not by Cognition).Jun 2024
- Open source: GitHub GiovanniDantonio (blog field giovannidantonio.com); his site lists 0PenAI (Zero Pen AI, camel_hack repo, 2025), eXpect: Predict the Truth (with repo under raghavpillai/eXpect, 2024), and a centralized Tor-inspired relay network with layered encryption (2024).2025
- Researcher at Harvard's Kempner Institute for AI 2024-2026: game theory and reinforcement learning in 2024, AI evaluations and benchmark manipulation in 2025; also a Teaching Fellow.Jan 2024
- Public headline says "Building Cognition; prev. thinky, deepmind, harvard, 2x phio gold".
- Writes long-form essays at giovannidantonio.com, including 'The science of benchmarking: from Zero to Hero [Part I]' (2026-08-12) and 'How to actually learn' (2026-02-18).Aug 12, 2026
- Giovanni D'Antonio is a coauthor of the paper 'SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge' with Lukas Haas, Gal Yona, Sasha Goldshtein and Dipanjan Das, published as a Tech Report in 2025.2025
- VC exposure 2024-2025: Bain Capital Ventures winter internship and Emergence Capital investment team (Spring 2025), per his LinkedIn stealth-startup entry.Sep 2024
Research Team + Office of the President
Helping people hire, sell, and change the world. Current sequoia scout. Companies: Sable, Sail Research, Universal Constructor, Intelligence.ai, MIRA, Tzafon, CompLabs, Pantheon, Creatorwood [...]
Trying to connect research <> product, promo in 3mo; Cooking on evals, RL, and a very, very cool product.
Researcher at the Kempner Institute for AI (Selection of papers in the "Projects" section) - 2024: Focus on Game Theory & Reinforcement Learning - 2025: Focus on AI Evaluations & Benchmark Manipulation Teaching Fellow - Spring 25: ES 94 with Prof. Josh Lerner - Entrepreneurship and Innovation: Case Study Perspectives - Fall 25: ES 30 with Prof. Spencer Rascoff (Co-Founder Zillow) - Startups: From Idea to Exit
Gemini Core Model Team, Frontier Evals
My main goal is to learn from both failures and successes to build something people really care about. - Bain Capital Ventures (BCV): Winter internship - Emergence Capital: Investment Team (Spring '25) - Human Capital: Investment Team (Fall '24)
Built a lot around AI Agents, Digital Twins, Cyberattacks, loved every moment of it, learned a ton. Company got acquired by depthfirst.com Check out at https://www.giovannidantonio.com/coding
Taught a full semester fellowship centered on Machine Learning to Harvard Students Topics Covered: Neural Networks CNNs + Computer Vision RAG + Prompt Engineering Transformers + Text Generation Diffusion + MultiModal Generation RL
Concurrent with my Bachelor, took more courses than expected because education is fun
To address complex modern challenges, leaders need innovation, problem-solving, and interdisciplinary skills. The Harvard Undergraduate Technology Innovation Fellows (HUTIF) Program, a collaboration by FAS, SEAS, and HBS, develops this new generation of tech entrepreneurs. Funded by the Roberts Family Foundation, the program selects 10-15 students (Roberts Family Fellows) from any concentration, helping them build the knowledge, skills, and community to launch and lead innovative endeavors.
Lots of evals flying around these days. Decision makers make decisions with numbers they do not truly understand. Hereโs a short starting guide to the science of benchmarking, so you can tell which numbers actually mean something!
Linkedin is a very different platform from X. Time to break the third wall, so you all know what is happening there. ๐ซก ๐ ๐ ๐
It might sound a bit cheesy, but as a new grad, life feels completely full of opportunity. Compared to where I was a few years ago, I feel incredibly lucky to be in a position where Iโm no longer just optimizing for "survival." My focus is gradually shifting from working just to get by, to finding true meaning in my daily life and career. Iโll be writing a longer piece on this reflection soon, but in the meantime, here is a look at what Iโve been working on at Thinking Machines Lab
All roads lead back to Google! https://lnkd.in/gxaSQ3JE
I spent a long time being jealous of people who had it easy. You know the type. The people who had the perfect internships handed to them. The ones who never seemed to struggle with rent or car trouble. The ones with the "smooth start." But lately, Iโm cheering my past. I see it every day: people who were protected from every minor fire while they were growing up. Because they never had to deal with small problems, they have no tools to deal with big ones. Now that they're in the real world, the first sign of a crisis feels like the end of the world. ๐๐ฆ๐ด๐ช๐ญ๐ช๐ฆ๐ฏ๐ค๐ฆ ๐ช๐ด ๐ข ๐ฎ๐ถ๐ด๐ค๐ญ๐ฆ ๐บ๐ฐ๐ถ ๐ค๐ข๐ฏโ๐ต ๐จ๐ณ๐ฐ๐ธ ๐ธ
Hire people who give a sh*t because there is something beautiful in working for work's sake. Growing up, I really struggled with this concept. Outside of this platform and outside of the bubble, most people are just trying to get to the end of the month. They want to minimize how much work they do. They try to spend as little time as possible in this thing called "work" so they can move on to that thing they call "life." The big change in my perspective came when I stopped seeing work as a way to survive and started seeing it as a way to live. Being free to leave a mark on the world through
A bit of a late (but BIG) life update: I officially joined Thinking Machines Lab as Member of Technical Staff! The energy here is surreal and I have met some of the kindest and most talented human beings. Weโre focused on taste and human-AI collaboration, not just raw scale. A lot is coming up, but it definitely feels that life (and my learning) is accelerating at unprecedented pace. Scary to not be in the classroom this semester, but deep down I was done and needed the challenge. Thank you to Mianna, Hannah, Erika, Soumith and Mira for making it happen. Time to get to work. Attached a p
Just move to silicon valley, please. The only case why you should not, is if you hate being surrounded by kids with massive dreams, pulling night shifts not because a boss told them to, but because they canโt sleep until they see their idea come to life. Itโs easy to be cynical about tech, but the inventions that came out of this 50-mile radius changed the trajectory of my life. And probably yours, too. Thank you to the people who are building the future. There is this wild, shared delusion here that we can actually fix things. That if enough smart people get in a room, the world might act
Young people need a framework of how to achieve โlearningโ. Everyone who's done anything meaningful in life tells young people some version of the same thing: at first, go where you'll learn the most. And it's good advice. But it's also strangely useless and not actionable. So I spent some time defining a framework to clearly decide which choices to make in order to maximize learning.
Maybe you're one of the lucky few who knew their passion at five years old. I wasn't, and took me a while to realize it was ok. I created my own server, started making money from it, and wanted to become the best gamer ever. I was 100 percent convinced my future was either going pro or becoming a YouTuber. But hereโs the thing. I didn't actually have some "divine calling" for those things. I just really liked playing the game. If I had followed that feeling without building any real world skills, Iโ
I finished 70 marathons in 2025. I am a big fan of consistency because I am not good enough to do things all at once. When people hearย "70 marathons" , it sounds like a massive, athletic feat. But the reality is much more mundane: I just walked 10,000 steps a day. My professor of software engineering taught me an important lesson. ๐๐ณ ๐๐ผ๐บ๐ฒ๐ผ๐ป๐ฒ ๐ผ๐ป ๐๐ต๐ฒ ๐๐ฒ๐ฎ๐บ ๐ป๐ฒ๐ฒ๐ฑ๐ ๐๐ผ ๐บ๐ฎ๐ธ๐ฒ ๐ฎ๐ป โ๐ต๐ฒ๐ฟ๐ผ๐ถ๐ฐโ ๐ฒ๐ณ๐ณ๐ผ๐ฟ๐, ๐ถ๐ ๐บ๐ฒ๐ฎ๐ป๐ ๐๐ต๐ฎ๐ ๐๐ต๐ฒ ๐๐ฒ๐ฎ๐บ ๐๐ฎ๐ ๐ฝ๐ผ๐ผ๐ฟ๐น๐ ๐ผ๐ฟ๐ด๐ฎ๐ป๐ถ๐๐ฒ๐ฑ, ๐ฎ๐ป๐ฑ ๐ฝ๐ฒ๐ผ๐ฝ๐น๐ฒ ๐๐ฒ๐ฟ๐ฒ ๐ผ๐๐ ๐ผ๐ณ ๐ฝ๐ผ๐๐ถ๐๐ถ๐ผ๐ป. In a perfect, smooth game, no one should have to be a hero. Most people think success comes from a single "heroic" effort or
For many, going home is a happy rite of passage. But for people who abruptly changed everything about their lives, it is a weird sensation. Rationality is not a choice anymore. Journaling becomes your only real home. I used to hate Naples. I actually told other kids I wasn't from there. I remember hiding in a closet just to avoid the subtle racism implied in the division between North and South. That adolescent boy wanted to change everything. His body, his mind, his heart. Then, I learned to love Naples. I learned to love the good sides and be realistic about the bad ones. Then, I became