Antoni Rosinol
@RosinolToni
Followers
2,117
Following
982
Media
65
Statuses
363
Co-Founder & Customer Support @ Stack AI (YC W23) | PhD @MIT | Enterprise-Grade AI Applications |
San Francisco, CA
Joined January 2016
Don't wanna be here?
Send us removal request.
Pinned Tweet
Big news from
@StackAI_HQ
! We closed a $3M Seed Round led by Google's AI fund
@GradientVC
.
We now power the GenAI agenda of public companies and fast-growing startups, have over 40k registered users, and support more than 200 organizations.
Our platform allows you to deploy
17
11
90
A small update on the history of SLAM, and its future :) Last year, NeRFs sparked a new and ongoing era of geo/photometric SLAM!
What lies ahead?
I think some exciting avenues will leverage LLMs, diffusion models, or use light fields as maps!
Same problem, new tools🚀
15
113
431
Volumetrically fusing the depths from Dense Monocular SLAM leads to poor reconstructions... EXCEPT, if you weigh the depths by their marginal covariance (inverse of information matrix). If you do so the reconstruction is up to 90% more accurate! See our WACV-accepted paper below.
3
30
244
While OpenAI's ChatGPT is down... Feel free to use: !
10
15
175
🚀 Great chat with Sam Altman today!
We discussed the future of
@OpenAI
, what lies ahead in the short and long term, and how
@StackAI_HQ
is becoming the middle layer of AI.
Sam was also super open to feedback on some of the current and future features from Open AI, and while I
7
6
80
Sigma-Fusion: Probabilistic Volumetric Fusion for Dense Monocular SLAM
3
10
64
Excited to present our latest work: "3D Dynamic Scene Graphs: Actionable Spatial Perception with Places, Objects, and Humans". The time is ripe for Spatial PerceptIon eNgines (SPINs) (
#SpatialAI
)!
Video:
Paper:
0
7
59
A holistic paper on Kimera, and how to bridge SLAM and Spatial Perception.
Some Gems:
- Kimera-PGMO: Simultaneous Pose-Graph and Mesh Optimization.
- Hierarchical Semantic Path-planning on 3D Scene Graphs.
- New open-source dataset: uHumans2 with 12 new scenes.
- Real-life DSGs!
3
11
59
Great article by
@khulick
! It reminds us how effortless tasks for humans are incredibly difficult for robots:
-
@Ken_Goldberg
touches on grasping
-
@tianminshu
explores common sense
- Ashley explains DALL·E
- And I go over robot navigation.
1
16
54
I had the pleasure of presenting our newest work on "Kimera: from SLAM to Spatial Perception", and I thought of sharing the talk with you:
The talk was given to the Robotics and Perception Group () led by
@davsca1
.
1
4
41
Woohoo! We are live! If you are looking to work with an amazing team from MIT, let me know! We are looking for the best developers out there 🚀
🔥Stack AI is now backed by
@ycombinator
! 🔥
We are building software to automate data labeling, detect data errors, and fine-tune Large Language Models (LLMs) for any application. Our solution streamlines ML teams' data fine-tuning process, saving thousands of hours.
3
7
33
3
4
33
It was a great pleasure to present Kimera + 3D DSGs, and my thoughts about the Future of SLAM !!
i.e. Tightly-coupling learning and SLAM + Differentiable Rendering.
Thank you
@Maani
for the invitation! Btw, Maani’s lectures are a gem :)
0
1
30
Stack AI went from an idea to a paid MVP in 14 days...
How did we achieve this?
a) Going from idea to MVP many times before, we got really good at this.
b) Having a stellar team that rows hard and in sync towards the same goal.
c) Talking to customers every day, and joining Y
3
4
24
Yeah! We are live at NYC Times Square :)
Thx
@brexHQ
&
@ycombinator
!
🚀 We opened access to our platform, and are now onboarding customers on a first-come-first-serve basis!!
🤖 Build and deploy powerful apps with Large Language Models
#llms
with us
@StackAI_HQ
3
2
26
We're thrilled to welcome
@kaihenthoiwane
to
@StackAI_HQ
! 🎉 🎉 🎉
🎨 Kai is joining us as a Design Engineer in our San Francisco office, bringing his design and web development expertise to the mix!
🧙🏽♂️ Being a Figma, React, and Next.js wizard, Kai will be leading our
4
1
25
Experimenting with Llama2 right now, and I'm in shock: it does follow instructions really well, here is a screenshot of the flow I built with Stack AI. Looking great!
0
2
20
🎉 We were selected to join the new
@AWSstartups
GenAI cohort!! 🎉
Having AWS on our side, will help us scale seamlessly our operations, while delivering an even better experience to our customers.
🚀Stack AI’s no-code platform helps businesses of all sizes deploy AI
We are excited to reveal the startups accepted to the
#AWSGenAIAccelerator
. These founders will prove what’s possible with technical help, mentorship, up to $300,000 in AWS credits & more.
Meet the startups solving the biggest problems with the power AWS:
2
15
38
3
2
20
Super excited to be part of YC W23 !!
Welcome to YC W23,
@mforstell
,
@bernaceituno
,
@rosinoltoni
, and team
@StackAI_HQ
!
Stack AI saves ML teams thousands of hours by streamlining their data labeling process. Check it out:
4
3
23
1
0
19
🚨A Cautionary Note to
@OpenAI
GPT Developers:
Be mindful that the raw data used in your chatbots might be more accessible than intended to those interacting with it.
I was able to access data from
@Levelsfyi
's GPT, which included comprehensive details on base salaries, stock
2
3
18
🎉🎉🎉New feature announcements (06/26)
📸YouTube Node: transcribe videos and use the text for further processing with LLMs, such as building summaries of videos!
🐍Python Node: You can now run arbitrary code on a Python Node!
👀Optical Character Recognition (OCR): Parse text in
1
1
19
Presenting at the Enterprise Gen AI Summit today!
Together with a an incredible lineup of speakers like HubSpots CEO,
@yaminirangan
, the CEO of Snowflake,
@RamaswmySridhar
, in addition to some of the brightest minds in the AI field from AWS, Zoom, ServiceNow, Perplexity AI,
1
2
17
📣📣📣 New feature announcements from
@StackAI_HQ
📈 Code interpreter plots: with Stack AI, you can now ask LLMs to generate base64 plots using plt and plotly. LLMs can now build nicely formatted reports from your spreadsheets 🤩
📋 Conversational Form Plugin: LLMs can run
1
4
16
GPT4 getting some intense competition 🤯
3 new models from Anthropic, 3 new benchmarks for coding, math, reasoning, language…
- Claude 3 Opus
- Claude 3 Sonnet
- Claude 3 Haiku
OpenAI seems to release new models when they get threaten, this might as well mean the release
Today, we're announcing Claude 3, our next generation of AI models.
The three state-of-the-art models—Claude 3 Opus, Claude 3 Sonnet, and Claude 3 Haiku—set new industry benchmarks across reasoning, math, coding, multilingual understanding, and vision.
570
2K
10K
1
2
15
Hot take: terrific Frontend Engineers are even more difficult to find than AI/ML engineers these days…
10
1
17
🎉 I'm thrilled to have
@IntelSoftware
by our side, backing us
@StackAI_HQ
, as we embark on our mission to ensure that LLM-powered tools are within everyone's reach. Our goal is to provide an end-to-end low-code drag-and-drop platform that any company can use.
Why not come
Welcome our newest
#IntelLiftoff
member,
@StackAI_HQ
! 🎉 They're building custom chatbots & search engines with Large Language Models in just minutes! Learn more about our program:
#AI
#Developers
0
2
9
2
2
16
A healthcare company spent **thousands** of hours manually assessing the quality of their calls — painstakingly evaluating calls against a long list of QA questions.
But that’s now a thing of the past.
With advancements in speech-to-text, models like
@DeepgramAI
Nova 2 Medical
5
3
16
0
3
14
We are not living in the future quite yet…
a) Waymo missed a right exit today
b) Our chatbots don’t consistently pass the Turing test
That said, having worked in both fields, progress in the last two years has been absolutely insane, and even a mildly conservative
0
0
14
🌟
@StackAI_HQ
hits Times Square 🌟
We are working relentlessly for AI to get your job... done! ✅
With our no-code platform, you can automate a TON of your business workflows; here are some examples (link to demos below):
- 🥇 Surfacing CVs to find talent for a new project.
1
4
15
Accepted paper to RSS 2020!
#SpatialAI
🚀🚀🚀
Fun fact: this paper would not be possible without the great
@MITAeroAstro
committee (
@SertacKaraman
et al.), who made the wise decision of ending the Qualification Exams' legacy. Free your students! 🙂
3
1
14
3
2
14
🤯Since day one, we received so much demand to help draft RFPs with LLMs.
However, ensuring the quality of an AI-generated document is a challenge. Companies need to enrich language models with their knowledge from several sources:
📚 Previous proposals
🏗 Previous projects
1
2
14
What a great success story,
@yummyridesvzla
@metavarce
!! Keep up the good work, and we are always here to help :)
@RosinolToni
Built on
@StackAI_HQ
crack!! Gracias por el gran apoyo
@RosinolToni
.
Aquí les dejo una de
@BernAceituno
y yo trabajando en lanzar esto al mercado.
0
3
15
0
0
13
🎉 It has never been easier to deploy an AI chatbot with WhatsApp. With
@StackAI_HQ
, you can deploy your LLM-powered bot to WhatsApp in 5 minutes.
0
1
13
Join us at NeurIPS to learn how to learn on 3D Meshes, manifold or not! (+ Francesco Milano,
@antoniloq
,
@davsca1
,
@lucacarlone1
) Link in comments!
- Paper:
- Code:
- Video:
1
1
13
🚀 Exciting News! OpenAI's GPT-4o Mini now available in
@StackAI_HQ
!
OpenAI's latest model, GPT-4o Mini, is smarter and more affordable than ever:
💡 Higher intelligence: Scores 82% on MMLU benchmark
💸 Cost-effective: Input tokens at $0.15/million, output tokens at
1
2
13
It was great to talk about my entrepreneurial journey with fellow LaCaixa scholars.
Coming from a purely academic background, it may feel daunting to jump head first into building a company… if you are about to graduate and are still unsure whether to do the jump, I’d be more
¿Estás pensando en emprender, pero no sabes por dónde empezar? 🤔
🤩 No te pierdas los 5⃣
#TipsForTops
del
#becariosFundlaCaixa
@RosinolToni
, doctorado en el
@MIT
y cofundador de
@StackAI_HQ
.👇
#D
íaMundialEmprendimiento
0
0
4
2
2
13
Happy to finally open-source Kimera! A c++ library for real-time localization, mapping, and semantic understanding using only a camera and an IMU Kudos to the amazing team: Marcus Abate, Yun Chang,
@lucacarlone1
#SLAM
#semantics
#robotics
Video:
1
4
13
📣📣📣 New feature announcements
💥Formless Form: Now you can collect user data without using tedious and boring forms. Instead, use Stack AI to build a Conversational Form to collect data using natural language.
🤖Llama-2: Access the most powerful open-source LLMs in Stack AI,
1
4
12
📣📣📣 New feature announcements
✉️**Gmail Integration**: you can now send and receive emails, and build chat like experiences with gmail!
🚀**Stack AI's Code Interpreter**: you can now have an LLM run Python Code in our own sandboxed, firewalled execution environment, which
0
3
12
✨🤩 New looks!
@StackAI_HQ
is the de-facto AI workflow automation builder for many enterprises now, and, as such, we are re-branding a bit to show our efforts towards becoming the go-to Enterprise-ready AI development platform out there.
🚀 Join us on our journey!
We'll be
0
3
12
I think there is no better way to test mobile responsiveness… 😅
0
1
11
What a great honor to be featured in Prof.
@fdellaert
presentation for his, and Prof.
@michaelkaess
, Test of Time award at RSS!
I just uploaded a talk explaining 3D Dynamic Scene Graphs, and how they enable Semantic Path Planning queries.
0
0
12
2
0
12
@AlphaSignalAI
@jon_barron
It would be great to see if it recognizes that this is a reprojection error loss with regularization.
3
0
9
Nice summary of YC’s W23 GenAI startups, featuring of course
@StackAI_HQ
.
🌟🔥We opened up our new no-code drag-and-drop composability tool for LLMs. Give it a try at:
Now you can automate arbitrarily complex business processes using the power of
2. Stack AI
Focus: Build LLM applications (i.e. ChatGPT) in minutes.
Location: San Francisco, CA.
Website:
4
16
295
0
2
10
🤯Is this the end of ?
I was recently welcomed by Meta / WhatsApp to join their waitlist for a new AI chatbot experience.
Haven’t tried it yet, but given how addictive character AI is, I can only imagine how much usage whatsapp will get out of this…
1
3
10
📣📣📣 New Stack AI features
🐍 **Code Interpreter to analyze CSVs:** Now everyone can analyze CSVs or Excel tables with Stack AI :) Simply add a table analyzer node and let the LLM use ***code execution*** to analyze and draw insights from it + you get plots of your data!
📻
1
3
11
📣📣📣 New feature announcements
@StackAI_HQ
📨 Gmail Receiver: Now, you can easily set up your Stack AI projects to receive emails and respond automatically. Add a Gmail Receiver, and your project will be automatically executed with each new unread email in your inbox 🤯.
📩
0
3
11
📣📣📣 New feature announcements
@StackAI_HQ
🧠 Shared LLM memory: Want to share memory across your LLMs? Leverage the "Shared Memory" node (currently in Beta) in the Utils section to share memory across your workflow's LLMs. Simply connect it as an input to an LLM and
3
3
11
Funny how so many people are like, "Well, should we move everything to
@AnthropicAI
now??"
If you had based your entire strategy on
@OpenAI
dominating for decades, you'd be facing a major dilemma right now.
Fortunately, our customers don't have to worry about that. We've
0
0
11
📣📣📣 New feature announcements
📸 **Text to Image: ** Turn plain text into captivating visuals! Introducing OpenAI DALL E-2 model, which, combined with LLMs, provides magnificent content.
💳 **HubSpot and Salesforce:** Connect your CRM to an LLM using Stack AI for engaging
1
2
11
@antoniloq
There is a Spanish saying "de broma en broma, la verdad asoma", which translates to something like: "Between joke and joke, the truth appears" 😂
0
0
11
What a journey!
Where it all began! Our co-founders
@RosinolToni
and
@BernAceituno
revisited the house they called home during
@ycombinator
after finishing their phd at
@MIT
. Nostalgic but full of energy for the next chapters in
#GenerativeAI
. Loving every bit of the journey!
#YCombinator
3
3
39
0
1
11
We already have access to LLaMA2 🦙
@StackAI_HQ
and we are amazed at how good it is! Getting close to GPT3.5 level of performance.
Open-source LLMs are a big deal. Imagine using AI right where your data is, keeping it private and safe.
That's our future – making AI work harder
We believe an open approach is the right one for the development of today's Al models.
Today, we’re releasing Llama 2, the next generation of Meta’s open source Large Language Model, available for free for research & commercial use.
Details ➡️
180
756
3K
0
2
10
🎉Best email I've received today!
Sent automatically from Stack AI using a flow involving Anthropic's Claude-2 LLM.
Beautiful 🤩
@StackAI_HQ
@BernAceituno
#LLM
#LLMs
#Anthropic
#StackAI
#ChatGPT
#OpenAI
2
1
9
⚡️💡 Moving at 299 792 458 m/s 💡⚡️
🕐🕐🕐
Just 5 hours after the OpenAI DevDay, we rolled out support for:
🚀 GPT-4 Turbo.
📋 JSON mode.
🖼️ GPT-4 with Vision.
🎨 Dall-E 3.
Which feature will you try first? Start stacking with our new updates today!
8
9
54
0
0
9
Beyond grateful for being a LaCaixa Fellow…
Our tribute to the kind of person you are |
#laCaixaFoundFellows
@BecarisFLC
0
1
10
Code:
Paper:
0
1
10
Best present we could ask for December, thank you
@sundarpichai
and team!
If you'd like to build your own custom ChatGEMINI.
Reach out to us
@StackAI_HQ
Talk soon!
#Gemini
#GPT
#GPTs
#AI
#OpenAI
#Google
#LLMs
Seeing some qs on what Gemini *is* (beyond the zodiac :). Best way to understand Gemini’s underlying amazing capabilities is to see them in action, take a look ⬇️
1K
7K
33K
0
4
10
❤️I love the new GPT-4 Turbo model, gpt-4-turbo-2024-04-09; it writes so much more naturally than the previous verbose and extra formal models... (see example below!)
👀The model also supports vision, where the model can “see” and interpret the content of images. Additionally,
3
2
9
@BernAceituno
gave an excellent presentation of
@StackAI_HQ
capabilities during
@AWSstartups
' Generative AI program. 🚀
Check out
@StackAI_HQ
No-code AI applications for all
Rapidly deploy chatbots, content, and customer support, making business operations smoother
@AWSstartups
2
1
6
0
3
9
@robo2902
SLAM++ and Kimera are object/semantic SLAM. I’m missing iLabel as an example of semantic SLAM using NeRFs though!
1
0
8
🤖📈 After experimenting with LLM-based eval, some thoughts:
💡Label-free eval — using LLMs to evaluate LLMs — is promising, easier for users to write assertions, and unit-tests.
🔥Costs/Latency is still a big concern, evaluation requires calls for each datapoint.
⚡️Label-free
1
2
8
Just arrived at the
#llama3hackathon
!
Great to see so many familiar faces :)
Como say hi to the
@StackAI_HQ
team if you are around, we are sitting at the
@LambdaLabs
booth 😅
@cerebral_valley
#llama3hackathon
0
0
9
I couldn't be more proud of having Manuel join StackAI as an AI/ML Engineer.
We are going to ship something truly amazing together, and with his talent and expertise, it is only a matter of time before we all get to enjoy a terrific platform to build and deploy AI-powered
🎉 Welcome to the team, Manu!
As our customer base expands, our mission is to continually enhance the core functionality of Stack AI, making it swifter, more resilient, and increasingly powerful. Manuel's expertise is key to achieving these goals.
His journey is nothing short
0
3
7
1
2
8
🎉 We are thrilled to welcome Maria de los Angeles to
@StackAI_HQ
!
Maria brings a ton of experience to the team, having previously worked at high-growth startups disrupting e-commerce, and also founding her own company operating throughout Argentina. Maria joins as a Front-End
0
1
8
@StackAI_HQ
we have built the infra and the UI/UX to quantitatively monitor the performance of isolated and chained LLMs to make sure that these are performing as expected in production and also to further fine-tune and improve the models.
0
2
6
🚀 Exciting News from StackAI! 🚀
We are thrilled to announce that Stack AI is now SOC 2 Type II and HIPAA compliant! 🎉
At Stack AI, the security and privacy of our customers' data are our top priorities. We are committed to maintaining the highest standards in data protection
1
0
8
🎉🎉🎉 1000 citations!
While I transitioned to the startup world (and what a ride that was!), I'll keep publishing research papers to help expand the boundaries of knowledge with
@StackAI_HQ
:)
0
1
6
Nothing more convenient than a React component!
You can now embed your
@StackAI_HQ
chatbot using
Check it out!
⚡️Integrate your
@StackAI_HQ
applications as React components!
The easiest way to integrate AI-based internal tools, search engines, assistants and many more.
🏗️ Contribute building your UIs and sharing with the community.
Check it out here:
0
1
8
0
2
7
🚨 Wow! Amazon 🤝 Anthropic. 🚨
Big moves for the AI world!
I have no doubt this will be positive for the world; with competition between (hopefully more than 2) companies, we will see faster progress, fairer outcomes, and broader accessibility.
Pumped to see where this
1
2
7
Falcon40B is now available as open source under the Apache2.0 license and free for commercial usage!
We'll be adding it really soon at
@StackAI_HQ
, follow us on Discord for more feature releases!
UAE's Falcon 40B, the world's top ranked open-source AI model from the Technology Innovation Institute (TII) has waived royalties on its use for commercial and research purposes.
#TII
#LLM
#FalconLLM
#Tech
#Innovation
#AI
#AbuDhabi
#UAE
37
253
1K
0
1
7
@levelsio
GPTs only give you so much...
...but if you want to:
1. 📩 Collect data (name, emails, ...) from your users
2. 🔗Release it under your website domain.
3. 📁 Ingest all your sources (website, docs, FAQs, Notion, Google Drive, etc)
5. 🤳Deploy it as a Landing Page
Any way I can make my GPT ask users for there email first? Then send it to my server with a function to save it?
47
5
192
0
3
7
Guess which one is from real data and which one is simulated?
3
0
7
1
1
6
🔥
@elevenlabsio
continues to impress! Our users are so happy with our ElevenLabs integration, which has allowed them to build super customized automated outbound sales pipelines, as well as many orher use cases, using voice 😍
Can’t wait to see what’s next!
@StackAI_HQ
#LLMs
Eleven Labs entering the dubbing game!
Convert your voice into other langages whist maintaining the same tone 🎉
No doubt they'll dominate this too!
Crazy fast feature shippers
1
1
7
1
0
6
I will be presenting our work on 3D Dynamic Scene Graphs in the Scene-Understanding workshop at CVPR 2020, everyone is free to join!
My talk is at 15:15 EDT (UTC-4:00), questions are more than welcome!
0
1
7
@OpenAI
@StackAI_HQ
@BernAceituno
Thank you as well to Steven Heidel for the great talk about OpenAI's most complex features, and Olivier Godement for offering us help with anything Open AI related. No wonder Open AI is doing so well with such a great team :)
0
0
6
🎉 We are sponsoring the Climate AI Hackathon
@StackAI_HQ
!!
Join us for an event bridging the gap between generative AI and real-world climate challenges, at the emblematic
#Shack15
!
0
1
6
Drag-and-drop your way to a fully deployed LLM-powered app:
Want to build an LLM chatbot in a few minutes? Here is how you can do it in Stack!
0
2
7
0
1
5
I found a way to know if you are talking to a top VC vs the rest…
The very best, and I mean Midas top 5, will just ask you one simple question…
1
0
6
So proud of my brother Gabriel Rosiñol!
Keep up the good work :)
0
0
6
New updates for Kimera-Semantics! Now using Voxblox' `fast` approach which is 10x faster than the `merged` approach with minor loss in accuracy particularly for small voxels: check it out !
0
1
6
@resend
is the best email platform for developers out there, give it a try! :)
Crafted to a level of detail that makes you 🤤
Generative AI is here to stay, and
@StackAI_HQ
is helping make that possible with a no-code platform.
Here's the story of how they are using Resend to handle their email communication.
1
0
16
0
0
6
Nice analysis from
@TheChowdhary
about
@ycombinator
's W23 batch.
Happy to see
@StackAI_HQ
ranking among one of the Top 10 YC W23 companies by monthly website visitors.
Top
@ycombinator
W23 startups by monthly web traffic
1.
@play_ht
2.
@resendlabs
3.
4.
@GlassHealthHQ
5.
@letsunifyai
6.
@rivet_gg
7.
@vocodehq
8.
@StackAI_HQ
9.
@CreatorML_
1
2
7
2
1
5
🎉🎉🎉
0
0
4
Sad but true!
PhD graduates in AI mostly take boring jobs at big tech companies due to short-term monetary incentives.
While understandable to some degree, it's also quite sad to see so many great researchers 'disappear' and give up their talent - join or do your own startup instead!
47
46
566
0
0
4
🌟 Kudos to
@perplexity_ai
for their Series B at a $1B valuation!
Hats off to Aravind and the team—your execution is stellar! 🚀
Perplexity's APIs are already in
@StackAI_HQ
, so you can give your custom ChatGPT state-of-the-art search capabilities (or shall I say your own
2
0
5
@ducha_aiki
I did work on a framework to have something like SuperSLAM, mostly a pytorch-enabled gtsam, but I am not continuing that project because it was dismissed as unnecessary given that industry has supposedly solved plain VO/VIO already… But hey if there is enough interest…
0
0
5