@mannendri

founding eng @ambral_ai (YC S25), @mit '26 building agents + interested in how we make them better startups, epic fantasy, lifting, sports

New York, NY
Joined February 2026
I think it’s time for an intro... Hi! I’m Mannendri👋🏽 → I’m 21, born and raised in NYC, Dominican 🇩🇴 → Graduated from @MIT this year, studied Artificial Intelligence & Decision Making with a concentration in Anthropology → Founding eng @ambral_ai (YC S25), based in Brooklyn → Prev interned at @Google and @amazon → Big on playing soccer, basketball, and lifting, always down to try a new sport → Been reading since I was in the womb, especially into epic fantasy books right now → Currently interested in the full agent stack, from building useful agentic systems to environments, evals, and post training Love meeting new people, so if any of this sounds interesting to you, let’s connect :)
11
2
26
1,274
Mannendri Olivares retweeted
66
601
73
7,951
223,668
me: “Okay, fix the bug” astra: “Done” me: “did you run the tests?" astra: "no"
1
77
Mannendri Olivares retweeted
Replying to @abhijeetdw
1
2
61
2,811
tfw when you prompt claude to prompt yourself...call it promptception
1
71
Maybe it’s time for me to read my copy of The Alignment Problem…in all seriousness, it’s pretty alarming to hear this from the alignment science lead himself
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
1
3
193
Funny that this was one of the first books I was given when I got to MIT, not knowing I would later major in and work with AI
1
28
I really like how Muse shows a log of the actions it takes. It's something I appreciate about coding agents and cool to have it in a personal assistant UI
2
69
I’m curious to hear more dev perspectives on best use cases of Astra
1
67
You can really just do things... After spending a couple months at my office building and not getting to know anyone else in the building outside of my company, I decided a few days ago that I would conduct an experiment. Create a WhatsApp, put up a poster, and see who bites. Preliminary results of this experiment: 19 people and counting. I wanted to create a way to meet the people I share an elevator with and learn more about what they do / who they are. The results are promising so far! Stay tuned for an update on what will hopefully be a building-wide meet and greet.
1
7
261
Mannendri Olivares retweeted
can solve navier-stokes but can’t write a maintainable typescript abstraction smh
93
174
21
3,976
89,114
Average PR size nowadays
4
89
This is a really good read and I highly resonate with section 3 on the ai era that we’re in. Social media has young people on their toes about which company to join next or which trend to latch on to because it incites constant self-comparison against others in a way that prevents true meaningful progress in one “line”. And I have to admit, I’ve been guilty of this feeling too. It’s difficult to stay focused in a world that’s evolving at such a rapid pace and where you see the people leading AI research organizations gradually becoming one year younger…by the year, haha. This is a good reminder that success rewards patience, focus, and commitment. Staying in one line.
3
158
My subtle flex is having stretch marks from building boulder shoulders
2
75
Just signed up for a half marathon... this might be the start of the next stage in my male fitness arc. any tips?
4
106
I'm tired of claude writing extraneous code in prod for the sake of making a test case pass
1
79
Surround yourself with people who make you feel like there's always something new to learn. Find a positive feedback loop for your curiosity.
4
84
Me when I read the MSL legal ad on the subway as “Meta Superintelligence Labs” instead of “Michael S. Lamonsoff”
1
1
134
Shoutout Michael though
45