GTM & DevRel @Lexsi_labs Community Projects Lead @Cohere_Labs
Bangalore
Joined August 2024
- Tweets32
- Following307
- Followers61
- Likes77
Ram Kadiyala retweeted
Today we are releasing our first version of MoE class model - OrionMoE-Small-v1.0 from Lexsi Labs.
It is the first MoE tabular foundation model to stand alongside frontier dense models, competing with models up to twice its size.
Tables are the most varied data enterprises run on, yet most foundation models still push every table through the same computation. We think the next step for tabular AI is models that adapt how they compute, not only how large they grow.
This is version 1. We will scale the OrionMoE family from here.
Early access: orionmoe.lexsi.ai
Technical report: arxiv.org/abs/2610.20418
Blog: lexsi.ai/resources/articles/…
Big kudos to Mohamed Bouadi, PhD, Nassim Bouarour, @AdityaTann18002, @SVinayKumar12 from our Paris lab.
Its experts sit at the core of the backbone, and each table chooses its own, routed by a learned summary of the dataset.
On TabArena it ranks 12th of 38 methods, above every tuned tree and neural baseline. On every suite, it rates above larger models.
#LexsiLabs #TabularFoundationalModels #OrionMoE #TFMs #EntepriseModels
Ram Kadiyala retweeted
🚀 Lexsi - AI Safety and Alignment Hackathon — $30,000+ in prizes and credits in collaboration with @Cohere_Labs and @huggingface
Build with Lexsi Labs' open-source post-training, safety and alignment stack, alongside Cohere Labs and Hugging Face.
🗓 Nov 1 – 21, 2026 · 🌐 Online · 👥 Teams of 1–3
Register here: hackathon.lexsi.ai/
If you are in Bangalore, you can also register for the hackathon kickoff event which covers invited talks from the organizing teams, deepmind, IISc, redhat etc...
🗓 Oct 31, 2026
Register here : luma.com/66jkhulx
ALT Hackathon poster for Lexsi Labs AI safety and alignment hackathon in collaboration with Huggingface and Cohere Labs
🚀 Lexsi - AI Safety and Alignment Hackathon — $30,000+ in prizes and credits in collaboration with @Cohere_Labs and @huggingface
Build with Lexsi Labs' open-source post-training, safety and alignment stack, alongside Cohere Labs and Hugging Face.
🗓 Nov 1 – 21, 2026 · 🌐 Online · 👥 Teams of 1–3
Register here: hackathon.lexsi.ai/
If you are in Bangalore, you can also register for the hackathon kickoff event which covers invited talks from the organizing teams, deepmind, IISc, redhat etc...
🗓 Oct 31, 2026
Register here : luma.com/66jkhulx
ALT Hackathon poster for Lexsi Labs AI safety and alignment hackathon in collaboration with Huggingface and Cohere Labs
Ram Kadiyala retweeted
A big announcement from Lexsi Labs:
Lexsi Labs (@Lexsi_labs) , Cohere Labs (@cohere ) and Hugging Face(@huggingface) are partnering on the Lexsi AI Safety and Alignment Hackathon, one of the biggest hackathon on post-training and AI Safety & Alignment.
Nov 1 to 21, 2026. Online, free, open to builders worldwide.
Teams build real post-training pipelines:
- Models: open-weights Tiny Aya and Aya Expanse from Cohere Labs
- Tooling: six open-source Lexsi Labs toolkits (AlignTune, CuratorKIT, AuditKIT, SafeTune, CircuitKIT, AgentTune)
- Artifacts: models, adapters, logs and datasets hosted on Hugging Face
Three tracks: fine-tuning and data curation, AI red teaming and safety, agentic fine-tuning.
Each submission chains at least two toolkits and reports before-and-after metrics.
$30,000+ in prizes and credits, GPU credits for up to 800 teams, Demo Day for the top 20, and interviews at Lexsi Labs for standout builders.
Kickoff: Bengaluru, Oct 31, with an online option.
Registration starts today at: hackathon.lexsi.ai
Thank you to @huggingface and @cohere for building this with us.
Ram Kadiyala retweeted
"Help help, the 70 IQ Indians are coming over here and beating us on all the standardized tests"
🤨
Ram Kadiyala retweeted
⚠️ IF THAT FLYDUBAI PLANE HAD CRASHED…
1️⃣ The Indian pilot would have been blamed instantly.
2️⃣ Racists would have flooded social media with “Indian pilots” narratives before investigators even reached the wreckage.
3️⃣ Pakistani troll accounts would have amplified it, turning a tragedy into another India-bashing opportunity.
4️⃣ AI171 and the Phuket incidents would have been dragged in as evidence of a “pattern of behaviour” of Indian pilots.
5️⃣ Every unrelated Indian aviation incident would suddenly have become part of a narrative about supposedly “unsafe Indian pilots”.
6️⃣ Meanwhile, Oman’s lack of any meaningful history of aviation terrorism would have been used to bolster the India pilot angle.
7️⃣ If Indian pilot Capt Smit hadn’t survived & fought back to save the plane, the world would have got an entirely different story.
Hiring a Growth Intern, preferably an undergrad with prior familiarity with post-training and/or safety and alignment.
app.screenloop.com/careers/a…
Ram Kadiyala retweeted
The wait is over! Meet Arduino VENTUNO Q, where AI takes action.
💬 Run local LLMs like Qwen 3, Gemma 4, and Qwen 3 VLM directly on the board
🧠 NPU + CPU + GPU +MCU: @Qualcomm Dragonwing IQ-8275 with up to 40 dense TOPS of AI performance and STM32H5 microcontroller for real-time control
🗄️ 16 GB RAM + 64 GB eMMC + expandable storage
🪁 Linux-powered, pre-loaded with Ubuntu OS + Zephyr RTOS
🛠️ Build AI faster using Arduino App Lab, @huggingface, @EdgeImpulse, and Qualcomm AI Hub
💨 Move seamlessly from prototype to production through Works with Arduino Certification Program
The first batch won’t last. Pre-order your VENTUNO Q with a free power supply and USB-C cable included! Get ready to enter a new era of Physical AI: arduino.cc/product-ventuno-q
Glad to see atleast one more person figure out the kind of shitpile we have in events & growth side of things in BLR. This is just one of many issues, would love to post more when there is enough evidence.
they’re buying attention by manufacturing a reason for people to discuss them, then eventually coming up with a clarification or an event that surprises everyone. half the people sharing the outrage tweets are probably from the same gang too. it might generate attention, but it’ll ultimately kill trust. they think they can fool people n that they’re being clever, but they’re actually being clever by half.
Ram Kadiyala retweeted
they’re buying attention by manufacturing a reason for people to discuss them, then eventually coming up with a clarification or an event that surprises everyone. half the people sharing the outrage tweets are probably from the same gang too. it might generate attention, but it’ll ultimately kill trust. they think they can fool people n that they’re being clever, but they’re actually being clever by half.
Ram Kadiyala retweeted
Have you signed up for the Cohere Labs Open Science Community ML Summer School? 📚 ✨
ML Summer School kicks off next week, join us to learn about ML Agents, Diffusion Models, Reinforcement Learning, Career Planning and so more!
Register today: cohere.link/xHeONbj
🎤 We're looking forward to welcoming @_amirbar, Elizabeth Lingg, @irombie, Julia Kreutzer, @KatrinaL2899, Laurie Voss, @maximelabonne, @itsNVA7, @PuyuanPeng, @_1024_m, @SidYaeger, @SreyanG, Tushar Udeshi, Vivian White, Yash Mali
💙 A huge thank you to our community leads for organizing this program @AhmadMustafaAn1
@Rafay_Mustafa1,@KanwalMehreen2,@PrahithaM, Anu Adesina, @VivekS100 and @rahul_narava!
Ram Kadiyala retweeted
Let's get studious. 🏫
This month, the Cohere Labs Open Science Community ML Summer School returns! Join us to be part of a global community exploring the future of ML and hear from speakers across the industry.
Register to save your spot today: cohere.link/xHeONbj
We're looking forward to welcoming @_amirbar, Elizabeth Lingg, @irombie, Julia Kreutzer, @KatrinaL2899, Laurie Voss, @maximelabonne, @itsNVA7, @PuyuanPeng , @_1024_m, @SidYaeger, @SreyanG, Tushar Udeshi, Vivian White, Yash Mali 🔥
A huge thank you to our community leads for organizing this program @AhmadMustafaAn1, @Rafay_Mustafa1, @KanwalMehreen2, @PrahithaM, Anu Adesina, @VivekS100, and @rahul_narava! ✨
Ram Kadiyala retweeted
“What’s my approval rating?”
- It’s bad Mr Prime Minister. It’s -43. It’s…it’s in the mud.
“Could it get lower?”
- I mean anything COULD happen but realistically-
“Kill the ponies.”
- Wha-
“The ponies. The cute little ponies. Kill them.”
- Sir, they’re endangered
“Fuck ‘em. Make the call.”
Exclusive from @oliver_wright
Dartmoor ponies could be subject to mass culling to reduce the impact on biodiversity after a controversial ruling by the government’s environmental quango
Natural England has demanded that all livestock grazing on the moor is reduced by about 75 per cent to protect other habitats, plants and species
The move looks set to result in the culling of up to nine in ten of the semi-wild ponies as farmers prioritise their own cattle and sheep to remain within Natural England’s limit to minimise the impact on their own livelihoods
Natural England argued that the move was necessary to protect the diversity of Dartmoor, which is a designated site of special scientific interest
However, the plan goes against a government commissioned review into the future of Dartmoor, published two years ago, which concluded that Natural England “should not take actions likely to result in a reduction in pony numbers”, adding they were “invaluable for conservation grazing”
The move has led to claims that the quango is acting as judge, jury and executioner of the ponies — a species which is itself seen as endangered
Dartmoor ponies could be put to death under biodiversity plans
thetimes.com/article/ba529f3…
We have a bi-weekly meets and are open to contributions. No prior exp or publications are needed. We also offer co-authorship to contributors.
🚀 Our community-led ML Agents group is kicking off a new collaborative project to build a Street Navigation Agent for more inclusive, region-aware local search. In many parts of the world, businesses exist physically — but not digitally.
They're exploring how AI can use tools like Google Street View to read storefront signs, apply distance & category constraints, and reason step-by-step to identify real-world services.
We’re also building a global benchmark across countries and languages to evaluate visual verification.
Will be there at the @ycombinator startup school. Open to discussing any potential matches/ideas on data/evals/agents.
Ram Kadiyala retweeted
Congrats to everyone involved in Kaleidoscope, a cross-institutional collaboration accepted to ICLR 2026 🔥
A special shoutout to @mziizm who championed this collaboration from day 1. It is the first accepted paper for many of the collaborators who are first time authors.
Ram Kadiyala retweeted
In 2025, our Open Science Community Leads showed what’s possible when AI research is built in the open.
38 leads, 17 programs, 125 guest speakers advancing open, collaborative AI across the world (find all talks here! cohere.link/uw1lvQQ). 🤯
Three of our papers have been accepted at AACL 2025 @aaclmeeting (2 Main, 1 Findings).
1. DSBC : Data Science task Benchmarking with Context engineering
arxiv.org/pdf/2507.23336
2. Uncovering Cultural Representation Disparities in Vision-Language Models
arxiv.org/pdf/2505.14729
3. Improving Multilingual Capabilities with Cultural and Local Knowledge in Large Language Models While Enhancing Native Performance
arxiv.org/pdf/2504.09753
Grateful to the co-authors
@SidYaeger @Siddartha_10 @jebish7 @delliott @alexrs95 @_sumand @_srishtiyadav @KanwalMehreen2
This was made possible through research grants from
@TraversaalAI @AnthropicAI @Cohere_Labs
(2/3)
Uncovering Cultural Representation Disparities in Vision-Language Models
arxiv.org/pdf/2505.14729
huggingface.co/datasets/Bias…
Key Highlights :
- We test several VLMs at country/culture recognition task in 3 settings : Open-ended, MCQs with similar or neighbouring countries, MCQs with random countries
- We also test them by image ablations (noise, rotations, greyscaling, etc..)
Findings :
- Country level biases do correlate with country wise availability of online data i.e more data or mentions >> less bias or misclassification. This contradicts the common assumption of western-favouritism.
- Image perturbations affect biases in a very random way even among models belonging to the same family.
- Language of prompt had negligible effect other than improving accuracy over countries that speak the language.
(3/3)
Improving Multilingual Capabilities with Cultural and Local Knowledge in Large Language Models While Enhancing Native Performance.
arxiv.org/pdf/2504.09753
A Hindi-English bi-lingual LLM with over 140 checkpoints trained with variations in data distributions.
Findings :
- LLM-translated data can work as good as real data to address lack of data
- Each task type has a different optimal data distribution amount, which could be determined by test runs on a subset of data.
- LLM-generated thinking texts were made descriptive yet concise, this led to less emission (less token consumptions) during evals for text-generation tasks while providing better performance.
Release :
- Open Data, Models and 140 Checkpoints
huggingface.co/collections/l…
huggingface.co/1-800-LLMs