Mstack AI’s cover photo
Mstack AI

Mstack AI

Chemical Manufacturing

Houston, Texas 5,868 followers

AI-Native Chemical R&D and manufacturing

About us

Mstack is building an AI-native platform transforming chemical R&D and supply. By unifying intelligence, chemical synthesis, and global manufacturing networks, we help innovators cut time-to-market by up to 90% and scale breakthrough products faster. Our mission: turn category complexity into competitive advantage through AI-driven innovation at global scale. We were founded to fix what’s broken in global chemical supply chains. Today, we support OEMs, formulators, and manufacturers across 15+ countries with a model that responds to real demand—not inventory guesswork. Here’s how we do it: • We partner with manufacturing sites across South Asia and the Middle East, avoiding over-reliance on China and staying insulated from tariffs, political instability, and supply shocks • We customize and sample quickly through our in-house synthesis lab • We ensure every shipment meets strict quality standards through independent lab testing • We provide supply chain visibility and proactive updates at every step • We cut costs without compromising on compliance or performance Our team is made up of experienced technical professionals from leading sectors including coatings, agrochemicals, water treatment, oil and gas, and personal care, pharmaceuticals, oleoresins and HI&I. They understand what matters in production and deliver exactly that. If you’re looking for flexibility, reliability, and a modern way to source specialty chemicals, Mstack is your advantage. Let’s move faster, with less friction.

Website
https://www.mstack.ai
Industry
Chemical Manufacturing
Company size
51-200 employees
Headquarters
Houston, Texas
Type
Privately Held

Locations

Employees at Mstack AI

Updates

  • View organization page for Mstack AI

    5,868 followers

    This week we unveiled the Chemstack AI platform at IKP Knowledge Park's "AI for Life Sciences" workshop, to a room of founders, scientists, and pharma leaders. It's now open to all, in early access. Presented by Ankit Guleria & Swayam Arora. One agentic platform to design, simulate, cost, optimize, and scale chemistry, with AI built in the operating system from day one. At its heart is the Virtual Reactor, simulating the complete reaction at quantum-mechanical accuracy, computing transition states, energy profiles, selectivity, and yield. It predicts the outcome before a single flask is touched. This isn't screening. It's first-principles physics. Working on synthesis R&D? See it run on your own chemistry. Reach out for early access, or explore at chemstack.ai #ChemistryAI #AgenticAI #VirtualReactor #DrugDiscovery #ProcessChemistry

    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
    • No alternative text description for this image
  • Mstack AI reposted this

    Mstack just had a paper accepted at ICML. In AI, there are many ways to measure progress. One of the clearest signals is when your work earns recognition from the world’s top researchers. International Conference on Machine Learning is one of the most influential machine learning conferences globally, with an extremely competitive acceptance process. Getting in is not just about building fast — it reflects real technical depth and meaningful innovation. It’s been incredible to watch Mstack AI and Shreyans Chopra go from strength to strength: global demand → an R&D-first culture → a native AI approach to accelerating research → and now this milestone at ICML. To us, this is a strong signal that the flywheel is accelerating. What excites us most is that they’re not only building impactful products today, but also contributing to the frontier of what AI can become tomorrow. Huge congratulations to the entire team - proud to be part of the journey.

    View profile for Raja Sekhar

    Director AI at Mstack Chemicals

    Thrilled to share that our paper "RETROSPECT: Proposal and Reranking for Single-Step Retrosynthesis" has been accepted at ICML 2026, AI for Science. Most single-step retrosynthesis systems collapse two distinct jobs — proposing disconnections and ranking them — into one stage. RETROSPECT pulls them apart and studies them as separate, reusable units. On the proposal side, we built the ChemAlign Transformer, an encoder-decoder SMILES model trained with hybrid root-aligned/random SMILES augmentation, Pre-LayerNorm, tied embeddings, EMA weights, and a Chemistry-Informed Neural Network loss (CINN). On the selection side, a LambdaMART reranker reorders the merged candidate pool using structural, reaction-template, upstream-score, and Quantum Descriptors. The practical takeaway is modularity. Because proposal and selection are decoupled, the ChemAlign Transformer can stand on its own or slot into larger ensemble-and-ranking pipelines, instead of being locked inside a single monolithic model. By proposing more accurate and more diverse disconnections, RETROSPECT can strengthen the multi-step route search that synthesis planning depends on — giving chemists better starting points when they work out how to actually make a molecule. We're actively hiring AI Researchers working at the intersection of AI/ML and Computational Chemistry — especially those excited about building chemistry-native AI systems, scientific reasoning models, and next-generation discovery platforms. Nice work done by Shreyas V, Ronit kumar choudhary, Arjun Verma and a thank you to the [ICML] Int'l Conference on Machine Learning AI for Science organizing team and reviewers. Deepak Warrier, Monish Kaul, Shreyans Chopra. #Retrosynthesis #AIforSciences #AIforChemistry #ComputerAidedSynthesis #ICML2026 #ICML2026 #DeepLearning #ComputationalChemistry #DrugDiscovery #Cheminformatics #AIResearch #Hiring

  • Mstack AI reposted this

    Delighted to share that we’ve had another paper accepted at the ICML AI for Science workshop. Retrosynthesis is a well-established research area, which makes meaningful novelty especially rewarding. One of the interesting aspects of our journey has been bringing perspectives and engineering discipline from the e-commerce and large-scale AI systems world into computational chemistry. Having 2 papers accepted from our research group at ICML AI for Science gives us confidence that we’re moving in the right direction. Chemistry remains an incredibly rich and compelling space for AI, and we believe this is only the beginning. Kudos to Raja Sekhar and the entire team for the tremendous effort behind this work. We’re actively hiring AI Researchers working at the intersection of AI/ML and Computational Chemistry - especially those excited about building chemistry-native AI systems, scientific reasoning models, and next-generation discovery platforms. #AIChemistry #ComputationalChemistry #AIForScience #ScientificAI #DeepLearning

    View profile for Raja Sekhar

    Director AI at Mstack Chemicals

    Thrilled to share that our paper "RETROSPECT: Proposal and Reranking for Single-Step Retrosynthesis" has been accepted at ICML 2026, AI for Science. Most single-step retrosynthesis systems collapse two distinct jobs — proposing disconnections and ranking them — into one stage. RETROSPECT pulls them apart and studies them as separate, reusable units. On the proposal side, we built the ChemAlign Transformer, an encoder-decoder SMILES model trained with hybrid root-aligned/random SMILES augmentation, Pre-LayerNorm, tied embeddings, EMA weights, and a Chemistry-Informed Neural Network loss (CINN). On the selection side, a LambdaMART reranker reorders the merged candidate pool using structural, reaction-template, upstream-score, and Quantum Descriptors. The practical takeaway is modularity. Because proposal and selection are decoupled, the ChemAlign Transformer can stand on its own or slot into larger ensemble-and-ranking pipelines, instead of being locked inside a single monolithic model. By proposing more accurate and more diverse disconnections, RETROSPECT can strengthen the multi-step route search that synthesis planning depends on — giving chemists better starting points when they work out how to actually make a molecule. We're actively hiring AI Researchers working at the intersection of AI/ML and Computational Chemistry — especially those excited about building chemistry-native AI systems, scientific reasoning models, and next-generation discovery platforms. Nice work done by Shreyas V, Ronit kumar choudhary, Arjun Verma and a thank you to the [ICML] Int'l Conference on Machine Learning AI for Science organizing team and reviewers. Deepak Warrier, Monish Kaul, Shreyans Chopra. #Retrosynthesis #AIforSciences #AIforChemistry #ComputerAidedSynthesis #ICML2026 #ICML2026 #DeepLearning #ComputationalChemistry #DrugDiscovery #Cheminformatics #AIResearch #Hiring

  • View organization page for Mstack AI

    5,868 followers

    Mstack is building "Chemstack AI" for every chemist on the planet, alongside Prof. Gautam Desiraju - one of modern chemistry's most influential minds. Prof. Desiraju is a founding figure in modern structural chemistry: 500+ papers, ~75,000 citations, h-index >105. Past President of the International Union of Crystallography (IUCr), 2026 Ewald Prize laureate - among the highest honors in the molecular sciences. His mentorship has been instrumental in shaping Chemstack AI - process chemistry software with atomic & functional depth across every reaction, route, and decision. Together, we're building what the world hasn't had: chemistry's first truly pharmaceutically correct corelation engine. Pictured: Prof. Desiraju with Deepak Warrier & Monish Kaul.

    • No alternative text description for this image
  • View organization page for Mstack AI

    5,868 followers

    One week out from the American Coatings Show! Mstack AI will be at Booth 607, May 5 to 7 at the Indiana Convention Center in Indianapolis. If you are working through a CASE sourcing challenge, qualifying a new resin or additive, or just want to see what we have been building, come by. We supply the chemicals and own the quality and delivery end to end. No middleman, no hand-offs. Calendar is filling up fast. DM or drop a comment to lock in a slot on the floor. #ACS2026 #ACS26Exhibitor #Coatings #CASE #SpecialtyChemicals Chris Fitzgerald Darin Shields Sam Kratzer Krishna Mohan Rai Joe Sentendrey Andross Corona

  • Mstack is heading to the American Coatings Show. Come find us at Booth 607. May 5 to 7, Indiana Convention Center, Indianapolis. Quick intro if we haven't met. Mstack is a specialty chemicals supplier, Houston-based and backed by Lightspeed. We cover the CASE market end to end: alkyds, acrylic emulsions, epoxy systems and curatives, monomers, polyols, UV-curables, and tackifiers. Here is what makes us different from a typical distributor. We supply the chemicals and own the quality and delivery. Our team qualifies the product at the source, manages the supply chain, and stands behind what lands on your dock. Not a broker passing orders through. If you are working through a sourcing problem, qualifying a new raw material, or just want to meet the team and talk shop, come by. DM us or drop a comment to lock in time on the floor. See you in Indy. Chris Fitzgerald Sam Kratzer Krishna Mohan Rai Joe Sentendrey Andross Corona Alonso #ACS2026 #ACS26Exhibitor #Coatings #CASE #SpecialtyChemicals

    • No alternative text description for this image
  • The last year building our CASE business at Mstack has been intense - in the best way possible. Tariffs. Supply shocks. Ongoing global disruptions. And yet, we didn’t slow down - we built through it. What this journey has reinforced for us is simple: resilient supply chains aren’t reactive, they’re engineered. In December 2023, we started with just one salesperson and a category lead. Today, we’re a fast-growing CASE team of 8 sales professionals, backed by a strong supply chain engine and a dedicated R&D function. We’re not just trading - we’re building. We’re not just supplying - we’re formulating. And if you love formulating as much as we do, you’ll understand what drives us. We’re now doubling down - investing in people, processes, and technology to build a world-class CASE business. And we’re hiring. Sales roles open. If you’re excited to build, solve, and grow in the coatings, adhesives, sealants, and elastomers space - let’s talk. This is just the beginning.

  • In chemistry, feasibility isn’t absolute: it’s volatile. Prices fluctuate, catalysts aren’t always available, environments differ across regions. What works in Boston may not work in Bangalore. With Chemstack AI, feasibility becomes locally practical. By embedding supply data, reaction parameters, and real-world availability into our knowledge graph, we enable chemists to discover what’s actually possible here and now.

    • No alternative text description for this image
  • Mstack AI reposted this

    The Indian manufacturing sector is being redefined by a new wave of startups, leveraging technology and global production networks to build a more resilient and efficient supply chain. We are proud to back visionary companies like ZETWERK, Mstack AI, Whizzo, and ZYOD who are raising the bar with their tech-driven solutions. From digitizing complex supply chains and enabling on-demand production to engineering cutting-edge materials, swipe through to learn how they are forging the future of India's manufacturing scene! Amrit Acharya Srinath Ramakkrushnan Shreyans Chopra Shrestha Kukreja Ankit Jaipuria Ritesh Khandelwal

Similar pages

Browse jobs