Latest news with #Gradient
Yahoo
21 hours ago
- Business
- Yahoo
Infisical raises $16 million Series A led by Elad Gil to safeguard secrets
Vlad Matsiiako is in the business of secrets. 'If secrets aren't there, then it's just not possible for software to run,' said Matsiiako, CEO and cofounder of Infisical. 'Databases can't connect to each other. Developers can't use any resources. AI agents can't integrate with the whole ecosystem. Basically, secrets are the glue that connects everything. And if they're not there, there's no way for organizations to secure systems. They can't really run things, it's not really operational. The majority of vulnerabilities these days are still related to secrets and identities.' Secrets refer to any and all sensitive information that protects and authorizes access to systems—think: passwords, encryption keys, API keys, and more. And Infisical is a secrets management platform for developers and companies, offering the tech to securely store, change, and retrieve vital credentials. Founded in 2022 by Matsiiako with Tony Dang and Maidul Islam, Infisical's now raised a $16 million Series A, led by Elad Gil. Y Combinator, Gradient, and Dynamic Fund participated, plus angels like Datadog CEO Olivier Pomel, Samsara CEO, and Valor CEO Antonio Gracias. Infisical's current customers include Hugging Face, Lucid, and LG. At first, Matsiiako expected he'd be mostly selling to tech companies, but found that non-tech-centric sectors were drawn to the product—banks, pharmaceutical companies, government, manufacturers, and mining organizations with extensive software infrastructure requiring secret management. Infisical has been expanding its product suite over the last year, to things like certificate management and SSH key management, and more. 'We're working with AI agents right now,' said Matsiiako, who's originally from Ukraine. 'Right now, the biggest growth area is all of these AI workloads, agents and applications—Cursor and so on. They write so much code, but they can only write code as quickly as they get access to different databases and different resources.' The secrets business is big—HashiCorp, an incumbent, sold to IBM for $6.4 billion in a deal finalized in February. And most reports show secrets management to be a growing market, worth well into the billions by 2030. In Infisical's case, the company said that it's cash flow positive and has grown revenue by 20 times over the last year. (The startup declined to disclose revenue.) 'As an investor and founder, I really look for things where there's a strong market pull,' said investor Elad Gil. 'You can be Sisyphus, rolling the boulder uphill all day. Or you can find something where the boulder's rolling downhill a little on its own—or at least, where the founder is pushing really hard, but gravity is helping. So, you're looking for the startup version of positive gravity…You want that market pull because without it, you're doing everything. And Infisical had that really clear product-market fit, that pull from customers who really wanted what they were building.' There were, of course, moments when the Infisical team did have to push boulders. Matsiiako, Dang, and Islam met as students at Cornell and cycled through other startup ideas—a VR marketplace, for one—before landing on secrets. They faced a time crunch in incorporating the company: They'd both just gotten into YC and as international students, they had visas to think about. The name 'Infisical' was born from that urgency, a blend of 'infinity' and 'physical'—meant to evoke something both expansive and tangible. And though the company has made fast progress—Infisical's software has been downloaded more than 40 million times globally in the past year—the early days were tough. When they entered Y Combinator's Winter 2023 batch, Infisical was a closed-source SaaS tool for managing developer secrets. The product worked technically, but the team struggled to gain traction and earn users' trust. A candid talk with YC group partner and managing director Dalton Caldwell changed everything. 'Obviously, there were [open source] YC success stories, like GitLab or Docker,' said Matsiiako. 'We were talking to Dalton, who'd worked with GitLab and others. We were like: We're hearing we need to open source this, but what if we lose our IP rights? And Dalton said: 'Guys, this thing is not working. What IP rights are you talking about? Just do it.'' The following week, the Infisical team open sourced the tech. It went viral, especially via Reddit. They woke up the next morning to users telling them that developer newsletters were featuring the project, a deluge of feature requests, and their first international contributor. 'Now, people could actually see the code,' said Matsiiako. 'They could see how the encryption works. And that was where trust came from.' As it turns out, secrets can only be shared with trust. IPO wrap…On Thursday, Circle went public on the NYSE, raising $1.05 billion. Today, Omada Health is slated to go public on the Nasdaq, targeting a $1 billion valuation. And not an IPO, but a 'wow': Anduril is now valued at $30.5 billion. Read my colleague Jessica Mathews' story here. See you Monday, Allie GarfinkleX: @agarfinksEmail: a deal for the Term Sheet newsletter here. Nina Ajemian curated the deals section of today's newsletter. Subscribe here. This story was originally featured on Sign in to access your portfolio
Yahoo
19-05-2025
- Health
- Yahoo
Gradient Denervation Technologies Announces FDA Breakthrough Designation for Pulmonary Artery Denervation System
PARIS, May 19, 2025 (GLOBE NEWSWIRE) -- Gradient Denervation Technologies announced today the company's pulmonary denervation system has received Breakthrough Device Designation by the U.S. Food and Drug Administration (FDA). This program creates an expedited review pathway for devices that have the potential to provide more effective treatment for life-threatening or debilitating conditions and meet FDA's rigorous standards for safety and efficacy. The Gradient Denervation System is a novel technology intended to treat patients with pulmonary hypertension and associated heart failure. Treatment is accomplished by ablating nerves around the pulmonary artery using therapeutic ultrasound energy in a minimally invasive, percutaneous procedure. The straightforward catheter platform was designed specifically for the pulmonary artery anatomy and leverages known interventional techniques. The treatment goal is to down-regulate the sympathetic activity in the pulmonary vascular tree to reduce vascular resistance and decrease pulmonary pressures. The PreVail-PH2 Early Feasibility Study is enrolling patients with pulmonary hypertension due to left-sided heart disease, classified by the World Health Organization (WHO) as Group 2 Pulmonary Hypertension. As many as two-thirds of heart failure patients around the world have elevated pulmonary vascular resistance, which is shown to lead to an increased risk of mortality and hospitalization. There are no approved drug or device therapies in the United States for this group of pulmonary hypertension patients. 'We are thrilled with this positive feedback from FDA. The granting of Breakthrough Device Designation marks another important milestone for our pulmonary denervation clinical development program,' said Martin Grasse, Chief Executive Officer of Gradient. 'We remain focused on completion of our early feasibility study as a crucial first step toward developing a targeted treatment option with the potential to improve outcomes and quality of life for these underserved patients.' About Gradient Denervation TechnologiesGradient Denervation Technologies is a Paris-based medical device company developing a minimally invasive, ultrasound-based device for the treatment of pulmonary hypertension. Gradient leverages intellectual property developed at Stanford University. The Gradient Denervation System is for investigational use only and is not approved for commercial use. For more information, please visit: CONTACT: Media Contact info@ in to access your portfolio
Business Insider
10-05-2025
- Business
- Business Insider
Check out the exclusive pitch deck that landed no-code AI agent startup StackAI a $16 million funding round from Lobby VC
The startup, StackAI, just raised a $16 million Series A funding round led by Lobby Capital. LifeX Ventures, Vercel CEO Guillermo Rauch, Weaviate CEO Bob Van, Gradient, startup accelerator Y Combinator, and Epakon Capital also participated in the round. Founded in 2022, San Francisco-based StackAI is a no-code platform for companies to develop AI agents that help with business functions. The startup's agents can interact with software such as Snowflake and Salesforce and be customized to complete back-office tasks like data entry, aggregating content, and categorizing information. StackAI was a member of YC's Winter 2023 batch and raised a $3 million seed funding round from Gradient, YC, Epakon Capital, Soma Capital, True Capital Ventures, and angel investors in April 2023. For cofounder Bernard Aceituno, one of the most surprising things about scaling StackAI has been the types of customers that have benefited most from the tech. "We found that sometimes the least technologically advanced companies — construction firms, local governments, and insurance — are the ones that gain the most value from AI agents," he told Business Insider. StackAI may be all about no- and low-code solutions, but Aceituno and his cofounder, Antoni Rosino, are coming at the problem from the opposite end of the spectrum. The pair met while earning their PhDs in computer science and artificial intelligence at MIT. They both graduated in 2022. Aceituno said that on StackAI's backend, the startup was leveraging AI itself to stay competitive as it grows, and as the tech evolves. "We heavily leverage AI for our development — Cursor and our own StackAI Agents to build 100-plus integrations and add new models as soon as they are announced," Aceituno said. AI agents are shaping up to be all the rage in Silicon Valley this year, with plenty of VCs showing a willingness to open their pocketbooks for startups that automate everything from sales calls, to data entry, to coding with AI. In the last month, Reco, which deployes AI cybersecurity agents, raised $25 million from Insight Partners; Artisan, which is replacing human employees with AI agents to complete repetitive tasks, raised $25 million from Glade Brook Capital; and Spur, which uses AI agents to debug websites, raised a $4.5 million seed round from First Round and Pear. Check out the 13-slide pitch deck StackAI used to raise its $13 million Series A funding round. StackAI pitch deck StackAI pitch deck
Business Insider
10-05-2025
- Business
- Business Insider
Check out the exclusive pitch deck that landed no-code AI agent startup StackAI a $16 million funding round from Lobby VC
Agentic AI continues to be a bright spot for VC investing in 2025, and one startup in the space just landed a fresh round of funding to bring no-code agents into the workplace. The startup, StackAI, just raised a $16 million Series A funding round led by Lobby Capital. LifeX Ventures, Vercel CEO Guillermo Rauch, Weaviate CEO Bob Van, Gradient, startup accelerator Y Combinator, and Epakon Capital also participated in the round. Founded in 2022, San Francisco-based StackAI is a no-code platform for companies to develop AI agents that help with business functions. The startup's agents can interact with software such as Snowflake and Salesforce and be customized to complete back-office tasks like data entry, aggregating content, and categorizing information. StackAI was a member of YC's Winter 2023 batch and raised a $3 million seed funding round from Gradient, YC, Epakon Capital, Soma Capital, True Capital Ventures, and angel investors in April 2023. For cofounder Bernard Aceituno, one of the most surprising things about scaling StackAI has been the types of customers that have benefited most from the tech. "We found that sometimes the least technologically advanced companies — construction firms, local governments, and insurance — are the ones that gain the most value from AI agents," he told Business Insider. StackAI may be all about no- and low-code solutions, but Aceituno and his cofounder, Antoni Rosino, are coming at the problem from the opposite end of the spectrum. The pair met while earning their PhDs in computer science and artificial intelligence at MIT. They both graduated in 2022. Aceituno said that on StackAI's backend, the startup was leveraging AI itself to stay competitive as it grows, and as the tech evolves. "We heavily leverage AI for our development — Cursor and our own StackAI Agents to build 100-plus integrations and add new models as soon as they are announced," Aceituno said. AI agents are shaping up to be all the rage in Silicon Valley this year, with plenty of VCs showing a willingness to open their pocketbooks for startups that automate everything from sales calls, to data entry, to coding with AI. In the last month, Reco, which deployes AI cybersecurity agents, raised $25 million from Insight Partners; Artisan, which is replacing human employees with AI agents to complete repetitive tasks, raised $25 million from Glade Brook Capital; and Spur, which uses AI agents to debug websites, raised a $4.5 million seed round from First Round and Pear. Check out the 13-slide pitch deck StackAI used to raise its $13 million Series A funding round. StackAI StackAI StackAI StackAI StackAI StackAI StackAI StackAI StackAI StackAI StackAI StackAI
Associated Press
02-05-2025
- Business
- Associated Press
MicroAlgo Inc. Develops Classifier Auto-Optimization Technology Based on Variational Quantum Algorithms, Accelerating the Advancement of Quantum Machine Learning
SHENZHEN, China, May 2, 2025 /PRNewswire/ -- MicroAlgo Inc. (the 'Company' or 'MicroAlgo') (NASDAQ: MLGO) announced today the launch of their latest classifier auto-optimization technology based on Variational Quantum Algorithms (VQA). This technology significantly reduces the complexity of parameter updates during training through deep optimization of the core circuit, markedly improving computational efficiency. Compared to other quantum classifiers, this optimized model has lower complexity and incorporates advanced regularization techniques, effectively preventing model overfitting and enhancing the classifier's generalization capability. The introduction of this technology marks a significant step forward in the practical application of quantum machine learning. Traditional quantum classifiers can theoretically leverage the advantages of quantum computing to accelerate machine learning tasks, but they still face numerous challenges in practical applications. Firstly, current mainstream quantum classifiers often require deep quantum circuits to achieve efficient feature mapping, which results in high optimization complexity for quantum parameters during training. Additionally, as the volume of training data increases, the computational load for parameter updates grows rapidly, leading to prolonged training times and impacting the model's practicality. MicroAlgo's classifier auto-optimization technology significantly reduces computational complexity through deep optimization of the core circuit. This approach improves upon two key aspects: circuit design and optimization algorithms. In terms of circuit design, the technology adopts a streamlined quantum circuit structure, reducing the number of quantum gates and thereby lowering the consumption of computational resources. On the optimization algorithm front, this classifier auto-optimization model employs an innovative parameter update strategy, making parameter adjustments more efficient and substantially accelerating training speed. In the training process of classifiers based on variational quantum algorithms (VQA), parameter optimization is one of the most critical steps. Generally, VQA classifiers rely on Parameterized Quantum Circuits (PQC), where updating each parameter requires computing gradients to adjust the circuit structure and minimize the loss function. However, the deeper the quantum circuit, the more complex the parameter space becomes, requiring optimization algorithms to perform more iterations to achieve convergence. Furthermore, uncertainties and noise in quantum measurements can also affect the training process, making it difficult for the model to optimize stably. Traditional optimization methods often employ strategies such as Stochastic Gradient Descent (SGD) or Variational Quantum Natural Gradient (VQNG) to find optimal parameters. However, these methods still face challenges such as high computational complexity, slow convergence rates, and a tendency to get trapped in local optima. Therefore, reducing the computational burden of parameter updates and improving training stability have become key factors in enhancing the performance of VQA classifiers. MicroAlgo's classifier auto-optimization technology, based on variational quantum algorithms, significantly reduces the computational complexity of parameter updates through deep optimization of the core circuit. It also incorporates innovative regularization techniques to enhance the stability and generalization capability of the training process. The core breakthroughs of this technology include the following aspects: Depth Optimization of Quantum Circuits to Reduce Computational Complexity: In traditional VQA classifier designs, the number of layers in the quantum circuit directly impacts computational complexity. To lower computational costs, MicroAlgo employs an Adaptive Circuit Pruning (ACP) method during optimization. This approach dynamically adjusts the circuit structure, eliminating redundant parameters while preserving the classifier's expressive power. As a result, the number of parameters required during training is significantly reduced, leading to a substantial decrease in computational complexity. Hamiltonian Transformation Optimization (HTO): Additionally, MicroAlgo introduces an optimization method based on Hamiltonian transformations. By altering the Hamiltonian representation of the variational quantum circuit, this technique shortens the search path within the parameter space, thereby improving optimization efficiency. Experimental results demonstrate that this method can reduce computational complexity by at least an order of magnitude while maintaining classification accuracy. Novel Regularization Strategy to Enhance Training Stability and Generalization Capability: In classical machine learning, regularization methods are widely used to prevent model overfitting. In the realm of quantum machine learning, MicroAlgo introduces a novel quantum regularization strategy called Quantum Entanglement Regularization (QER). This method dynamically adjusts the strength of quantum entanglement during training, preventing the model from overfitting the training data and thereby improving the classifier's generalization ability on unseen data. Additionally, an optimization strategy based on the Energy Landscape is incorporated, which adjusts the shape of the loss function during training. This enables the optimization algorithm to more quickly identify the global optimum, reducing the impact of local optima. Enhanced Noise Robustness for Real Quantum Computing Environments: Given that current Noisy Intermediate-Scale Quantum (NISQ) devices still exhibit significant noise levels, a model's noise resilience is critical. To improve the classifier's robustness, MicroAlgo proposes a technique based on Variational Quantum Error Correction (VQEC). This method actively learns noise patterns during training and adjusts circuit parameters to mitigate noise effects. This strategy markedly enhances the classifier's stability in noisy environments, making its performance on real quantum devices more reliable. MicroAlgo's classifier auto-optimization technology, based on variational quantum algorithms, successfully reduces the computational complexity of parameter updates through deep optimization of the core circuit and the introduction of novel regularization methods. This approach significantly boosts training speed and generalization capability. This breakthrough technology not only demonstrates its effectiveness in theory but also exhibits superior performance in simulation experiments, laying a crucial foundation for the advancement of quantum machine learning. As quantum computing hardware continues to advance, this technology will further expand its application domains in the future, accelerating the practical implementation of quantum intelligent computing and propelling quantum computing into a new stage of real-world utility. In an era where quantum computing and artificial intelligence converge, this innovation will undoubtedly serve as a significant milestone in advancing the frontiers of technology. About MicroAlgo Inc. MicroAlgo Inc. (the 'MicroAlgo'), a Cayman Islands exempted company, is dedicated to the development and application of bespoke central processing algorithms. MicroAlgo provides comprehensive solutions to customers by integrating central processing algorithms with software or hardware, or both, thereby helping them to increase the number of customers, improve end-user satisfaction, achieve direct cost savings, reduce power consumption, and achieve technical goals. The range of MicroAlgo's services includes algorithm optimization, accelerating computing power without the need for hardware upgrades, lightweight data processing, and data intelligence services. MicroAlgo's ability to efficiently deliver software and hardware optimization to customers through bespoke central processing algorithms serves as a driving force for MicroAlgo's long-term development. Forward-Looking Statements This press release contains statements that may constitute 'forward-looking statements.' Forward-looking statements are subject to numerous conditions, many of which are beyond the control of MicroAlgo, including those set forth in the Risk Factors section of MicroAlgo's periodic reports on Forms 10-K and 8-K filed with the SEC. Copies are available on the SEC's website, Words such as 'expect,' 'estimate,' 'project,' 'budget,' 'forecast,' 'anticipate,' 'intend,' 'plan,' 'may,' 'will,' 'could,' 'should,' 'believes,' 'predicts,' 'potential,' 'continue,' and similar expressions are intended to identify such forward-looking statements. These forward-looking statements include, without limitation, MicroAlgo's expectations with respect to future performance and anticipated financial impacts of the business transaction. MicroAlgo undertakes no obligation to update these statements for revisions or changes after the date of this release, except as may be required by law. View original content: SOURCE
