Rintu Kutum

@KCDH_A congratulates Dr. @ShraddhaKB, Assistant Professor (Research), KCDH-A and Dr. @rintukutum, Faculty Fellow & Data Scientist (Department of Computer Science), on securing a prestigious 𝐁𝐅𝐈-𝐁𝐈𝐎𝐌𝐄 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐆𝐫𝐚𝐧𝐭 for an 18-month research initiative.

I recently stepped away from delivering guest lectures at these IIT (n) x [Random Startup] collaboration courses a while ago. The model didn’t sit right with me… ₹30-40k for content that’s often outdated when the cohort even finishes, taught by people who haven’t shipped anything in healthcare AI or done any robust research themselves. I’ve been thinking about what would actually move the needle for Indian doctors, residents, and medical students trying to make sense of this space. Most are confused and I get requests every day from radiologists asking what is a good course for AI. But I do not want to offer another certificate to hang on LinkedIn (there are many on Coursera) but genuine literacy that clears doubts of clinicians, spot overhyped tools, and make informed career decisions as AI reshapes practice. So I’m putting together a small team to build something different: an affordable, modular course that meets people where they are in their AI journey… whether a final-year MBBS student curious about the field or a practicing radiologist trying to figure out AI. What am I looking for? Unit economics that work. No ₹40k overheads subsidizing fancy production. Enough to ensure that we are able to sustain and continue delivering some value. Will be starting with foundations. If it delivers value, we build forward with domain experts. If it doesn’t, we learn why. This is very personal for me as I gave up teaching when I left AIIMS to focus on research full-time… and I’ve missed it more than I expected. If you’re someone who’s passionate about teaching in this space and wants to build something meaningful, feel free to join. We may fail to compete with big names and associations and may also raise some eyebrows. But the attempt feels necessary.

Dr. Beena Pillai joined as the new Director: Dr. Beena Pillai, currently serving as Chief Scientist at CSIR-IGIB, New Delhi, joined as the new Director of BRIC-RGCB. Dr. Beena was accorded a warm welcome upon her arrival and was received by Dr. T. R. Santhosh Kumar, Director (Additional Charge); Mr. S. Mohanan Nair, Chief Controller; along with faculty and staff. An accomplished scientist in RNA biology and neuronal development and function, Dr. Beena’s lab at CSIR-IGIB studies the role of RNAs in early development, neurogenesis, adult behaviour, disease susceptibility, and miRNA markers for human diseases. @DrJitendraSingh @rajesh_gokhale @DBTIndia @IndiaDST @BricDbt @beena_tweets @IGIBSocial @ChandrabhasN

We’re open sourcing the first document OCR benchmark for the agentic era, ParseBench. Document parsing is the foundation of every AI agent that works with real-world files. ParseBench is a benchmark that measures parsing quality specifically for agent knowledge work: ✅ It optimizes for semantic correctness (instead of exact similarity) ✅ It has the most comprehensive distribution of real-world enterprise documents It contains ~2,000 human-verified enterprise document pages with 167,000+ test rules across five dimensions that matter most: tables, charts, content faithfulness, semantic formatting, and visual grounding. We benchmarked 14 known document parsers on ParseBench, from frontier/OSS VLMs to specialized parsers to LlamaParse. Here are some of our findings: 💡 Increasing compute budget yields diminishing returns - Gemini/gpt-5-mini/haiku gain 3-5 points from minimal to high thinking, at 4x the cost. 💡 Charts are the most polarizing dimension for evaluation. Most specialized parsers score below 6%, while some VLM-based parsers do a bit better. 💡 VLMs are great at visual understanding but terrible at layout extraction. GPT-5-mini/haiku score below 10% on our visual grounding task, all specialized parsers do much better. 💡 No method crushes all 5 dimensions at once, but LlamaParse achieves the highest overall score at 84.9%, and is the leader in 4 out of the 5 dimensions. This is by far the deepest technical work that we’ve published as a company. I would encourage you to start with our blog and explore our links to Hugging Face to GitHub. All the details are in our full 35-page (!!) ArXiv whitepaper. 🌐: Blog: llamaindex.ai/blog/parsebenc… 📄 Paper: arxiv.org/abs/2604.08538… 💻 Code: github.com/run-llama/Pars… 📊 Dataset: huggingface.co/datasets/llama… 🎥 YouTube: youtube.com/watch?v=g5p7G-…

1/ Annbatch unlocks terabyte-scale training of biological data in anndata 🚀 Check out our preprint here: arxiv.org/abs/2604.01949 📄 Grab the code on GitHub: github.com/scverse/annbat… Massive shoutout to project leads @felix_f0097 and Ilan Gold for driving this home! 🧵👇

PET scans are critical for tracking neurodegenerative disease, but results vary across scanners. This study shows AI can harmonize measurements across platforms, reducing bias by >80% and enabling more consistent diagnosis and monitoring. nature.com/articles/s4174…

"These findings highlight limitations of current pLMs for mutational effect prediction and suggest that dataset composition, rather than model architecture or training, is the primary driver of predictive success." Again and again and again. Exact same problem with DNALMs

@harvardchanirc is pleased to organize a 2-day workshop on "Improving Public Health Through Climate Action." 🗓: June 9-10, 2026 📍: Mumbai 🔗: hsph.me/climate-action… Deadline: April 12, 2026

Happy to share our recent work that took a long time in making! What do lipids DO to PROTEINS? Well, we show that lipids allosterically open the receptor binding pocket of LC3, a key protein in autophagy. Using this fundamental mechanism, we engineered LC3 proteins that are either functionally “inactivated”or “hyper-active”on autophagosomal membranes A novel lipid-triggered allosteric site modulates LC3-LIR receptor binding activity biorxiv.org/content/10.648…

We’re thrilled to open-source LabClaw — the Skill Operating Layer for LabOS by Stanford-Princeton Team One command turns any OpenClaw agent into a full AI Co-Scientist. Demo: labclaw-ai.github.io Dragon Shrimp Army reporting for duty 🦞🔬 #AIforScience #OpenClaw

🥳 Our 3rd most downloaded #REVIEW in 2025: Acquired resistance in cancer: towards targeted therapeutic strategies #therapyresistance #cancertherapy @NatResCancer #NRCtop10of2025 dlvr.it/TQvHtz

Cracking the case: how did adenoviral vector vaccine such as the J&J's Covid induce very rare and potentially fatal clotting, bleeding? An elegant proof of molecular mimicry—genetic background + rogue antibodies nejm.org/doi/full/10.10… theatlantic.com/science/2026/0… @rkhamsi @TheAtlantic science.org/content/articl… @kakape @GretchenVogel1 @ScienceMagazine

We give a glimpse at some of the capabilities of IsoDDE: - predicting novel biomolecular structures with 2-3x the accuracy of previous methods (including our own!) - the ability to predict binding affinity, one of the holy grail quantities of rational drug design, better even than physics simulations - the ability to highlight and uncover new pockets that had not previously been discovered 2/7

Today we share a technical report demonstrating how our drug design engine achieves a step-change in accuracy for predicting biomolecular structures, more than doubling the performance of AlphaFold 3 on key benchmarks and unlocking rational drug design even for examples it has never seen before. Head to the comments to read our blog.

Introducing gRNAde: our own little "AlphaGo Moment" for RNA design! 🧬🚀 📝: tinyurl.com/gRNAde-paper Unlike proteins, RNA design has long relied on "wisdom of the crowd" (human experts) or the slow crawl of directed evolution — gRNAde changes that! 🧵👇

Amazing paper from @PhilHolliger science.org/doi/10.1126/sc…

AI is cool and all... but a new paper in @ScienceMagazine kind of figured out the origin of life? The paper reports the discovery of a simple 45-nucleotide RNA molecule that can perfectly copy itself.

From lecture videos to national addresses, from textbooks to novels, Sarvam Studio is powering creation of multilingual content. We are excited to see what you will build. Read our blog for more details and to get started. sarvam.ai/blogs/sarvam-s…

The UN General Assembly has appointed 40 experts of the Independent International Scientific Panel on AI. 🏛️ This body begins its work as a scientifically-grounded foundation, ensuring global understanding is driven by evidence-based scientific assessments. #DigitalCooperation

Honoured to be appointed to the Independent International Scientific Panel on AI by the UN General Assembly. I look forward to working with an esteemed group of scientific leaders in AI on the panel's objectives.