<div>
  <ul>
    <li>
      <p>
        Scale AI, a data annotation and validation platform, has collaborated with OpenAI, a research lab and applied AI company, to provide fine-tuning tools for the latter's large language model, GPT-3.5.
      </p>
    </li>
    <li>
      <p>
        The partnership enables Scale's customers to prepare and utilize their proprietary data with Scale's Data Engine platform, allowing them to refine GPT-3.5 to create tailored models for their unique business requirements. These models will be able to reference and cite this data in its responses.
      </p>
    </li>
    <li>
      <p>
        Moreover, the fine-tuned models will be reviewed by Scale&rsquo;s human experts to ensure that it meets performance expectations and other safety and regulatory requirements.
      </p>
    </li>
  </ul>
  <p>
  </p>
</div>


<div>
  <p>
    <a href="https://scale.com/blog/open-ai-scale-partnership-gpt-3-5-fine-tuning">
      Press release
    </a>
  </p>
  <p>
    <a href="https://techcrunch.com/2023/08/24/openai-partners-with-scale-ai-to-allow-companies-to-fine-tune-gpt-3-5/">
      TechCrunch
    </a>
  </p>
</div>


Machine Learning Infrastructure

Data annotation and validation platform, Scale AI, has partnered with OpenAI to provide tools for fine-tuning the GPT-3.5 language model. This collaboration allows Scale's customers to use their own data to refine GPT-3.5 and create personalized models for their specific business needs. The models will be able to refer to and cite this proprietary data in their responses. Additionally, Scale's human experts will review the fine-tuned models to ensure they meet performance expectations and comply with safety and regulatory requirements.

Scale AI and OpenAI collaborate to fine-tune GPT-3.5 with custom data

Empowering Tomorrow's Innovations with Generative AI Infrastructure

Generative AI Infrastructure

ANALYST_QUICK_TAKE

<div>
 <p>Snowflake, a company specializing in cloud-based data and AI infrastructure, has agreed to acquire Datavolo, a provider of multimodal data pipelines for AI, for an undisclosed sum.</p>
 <p>Datavolo uses Apache NiFi, an open-source project initiated by the NSA, to develop a platform that streamlines data flows between enterprise data sources. Datavolo's technology is designed to process, extract, clean, transform, and enrich data for GenAI applications.</p>
 <p>In the wake of the acquisition, Snowflake intends to integrate Datavolo into its own platform, thereby enhancing its data engineering capacities. This will allow customers to swap out complex, direct connectors for more flexible and reusable pipelines. Snowflake also plans to continue maintaining the Apache NiFi project, while also aiming to expand its presence in the public sector.</p>
</div> <p><b>Analyst QuickTake:<br>
</b> This strategic acquisition aligns with Snowflake's focus on the public sector and follows its previous acquisition of <a href="https:<br>
//sp-edge.com/updates/33819"> Night Shift Development </a> in September 2024. This move aimed to enhance decision-making for government agencies by establishing a dedicated entity for the US public sector, enabling new methods of collaboration with public organizations.</p>

Snowflake acquires Datavolo to enhance data integration capabilities for undisclosed sum

Data Infrastructure & Analytics

<div>
 <p>Snowflake, a company specializing in cloud-based data and AI infrastructure, has announced its intentions to acquire Datavolo, a firm that provides robust data pipelines for AI applications. The financial details of the agreement remain undisclosed.</p>
 <p>Datavolo makes use of Apache NiFi, an open-source project originally developed by the NSA, to design a platform that automates the movement of data between various enterprise data sources. Their technology handles processing, extraction, cleansing, transformation, and enrichment of data for GenAI applications.</p>
 <p>Post-acquisition, Snowflake intends to absorb Datavolo into its platform in a move to bolster its data engineering capabilities. This merger will enable Snowflake's customers to substitute complicated point-to-point connectors with versatile, reusable pipelines. Furthermore, Snowflake plans to continue the upkeep of the Apache NiFi project and looks forward to expediting its growth in the public sector.</p>
</div> <p><b>Analyst QuickTake:<br>
</b> This strategic acquisition aligns with Snowflake's focus on the public sector and follows its previous acquisition of <a href="https:<br>
//sp-edge.com/updates/33819"> Night Shift Development </a> in September 2024. This move aimed to enhance decision-making for government agencies by establishing a dedicated entity for the US public sector, enabling new methods of collaboration with public organizations.</p>

Conservation Tech

<div>
 <p>OceanWell, a forward-thinking desalination startup, has garnered USD 11 million in Series A funding. Prominent contributors include Kubota Corporation, Jon Hemingway's family office of Carrix Ports, and Charles McGarraugh, who previously headed Goldman Sachs Metals Trading.</p>
 <p>The investment is intended to support the scale-up of the company's revolutionary modular deep-sea water farm technology. An initial water farm pilot in cooperation with Las Virgenes Municipal Water District (LVMWD) in California is also on the cards. OceanWell's ambitious plan is to construct 15 water farms across the world.</p>
 <p>The novel technology developed by OceanWell harnesses the hydrostatic pressure at depths of 400 meters to facilitate reverse osmosis desalination. Each of the company's pods has the capacity to generate up to one million gallons of fresh water every day. This method provides a 40% reduction in energy usage compared to traditional desalination techniques, and it offers additional benefits in terms of marine life preservation.</p>
 </div>

OceanWell raises USD 11 million in Series A funding to build water farms

Foundation Models

PROMINENT_CATEGORY

<div>
 
 <p>H, a Paris-based AI startup created by former employees of Google, has introduced Runner H, a platform leveraging Artificial Intelligence for business automation.</p>
 <p>This platform, Runner H, is developed on a unique, compact LLM with 2 billion parameters. It is designed to address three primary use cases: RPA (Robotic Process Automation), quality assurance, and business process outsourcing. The design of the platform permits users to either choose from already built agents or design their own using API's.</p>
 
</div> <p><b>Analyst QuickTake:<br>
</b> This news follows its funding round of <a href="https:<br>
//sp-edge.com/updates/29664"> USD 220 million </a> in June this year to acquire datasets and compute power to advance artificial general intelligence (AGI). The company claims its compact model outperforms Anthropic's "Computer Use" by 29% based on WebVoyager benchmarks, as well as models from Mistral and Meta. H states the smaller parameter count makes the platform more efficient in terms of cost and operations.</p>

H launches Runner H, an AI agent for business automation

INCUMBENTS_PRESENT

<div>
 
 <p>Capgemini expands its Intelligent App Factory on Azure through a partnership with Mistral AI and Microsoft to boost GenAI adoption. This partnership will incorporate Mistral AI's language models into Microsoft Azure AI Studio. Furthermore, Capgemini's RAISE platform will be utilized to deliver customizable, cost-effective, and secure GenAI solutions. </p>
  
</div>

Capgemini partners with Mistral AI and Microsoft to expand GenAI solutions globally

<div>
 
 <p>Biotechnology AI solutions provider, Converge Bio, has introduced a platform to aid companies in integrating biology-focused large language models (LLMs) into their R&amp;D processes. They have secured USD 5.5 million in seed funding, led by TLV Partners.</p>
 <p>This funding will be used to grow their team, gain customers, publish scientific articles on antibody design, and create their own foundation model.</p>
 <p>The firm states that the platform provides data enrichment options for augmenting existing LLMs with particular biological data, fine-tuning features for specific research, and an explainability layer enabling researchers to comprehend model outputs in depth.&nbsp;</p>
 <p>GenAI models for drug discovery are provided by Converge Bio. It is reported that the platform aims to ease the application of AI in specialized fields such as microbiology and immunology, facilitating researchers to effectively use LLMs in their routine research workflows.&nbsp;</p>
 
</div>

Converge Bio launches biotech LLM platform with USD 5.5 million seed funding

<div>
 
 <p>Anthropic, an AI models developer, has joined forces with Snowflake, a data and AI infrastructure company based on cloud, to incorporate Claude 3.5 models into Snowflake's AI platform, Cortex, to bolster AI development capabilities.</p>
 <p>The alliance allows Snowflake clients to use Anthropic's Claude models in the secure environment of Snowflake without transferring data externally. Alongside, Snowflake plans to utilize Claude to run its GenAI agents both semi-autonomously and autonomously, with the pre-developed agents primed for Claude from the inception.</p>
 <p>Although this is the seventh LLM partnership for Snowflake, it stands out from the earlier ones as it involves the direct employment of Claude in its agency tools by Snowflake, with an emphasis on preserving security, governance, and privacy while enabling users to take advantage of GenAI abilities.</p>
 
</div>

Snowflake partners with Anthropic to integrate Claude AI models into Cortex AI platform

<div>
 
 <p>Chinese AI research firm, DeepSeek, has introduced a reasoning AI model called DeepSeek-R1. This model prides itself in self-fact-checking by dedicating ample time to examining questions or searches. According to the company, DeepSeek-R1 has the ability to forecast and perform a succession of tasks to reach a conclusion, taking even tens of seconds to scrutinize complicated inquiries. The model's performance competes with OpenAI's o1-preview model on AI benchmarks like AIME and MATH.</p>
 <p>The firm DeepSeek, established in 2023, pioneers in developing advanced large language models (LLMs) and coding assistants, integrating general language procession with specialized skills. The company's top-tier models include DeepSeek-V2.5, which is a general-purpose LLM.</p>
 
</div>

DeepSeek releases AI reasoning model DeepSeek-R1

Technology providers that enhance satellite operations

Next-gen Satellites

<div>
 <p>ICEYE, a company specializing in Earth observation, is to take the lead role in a consortium created to develop space technologies for the Finnish Ministry of Defense F-35 program. Other participants in the consortium will include Lockheed Martin, Insta, Huld, DA-Group, and the Finnish Meteorological Institute.</p>
 <p>The objective of this partnership is progress in the field of space and joint intelligence technologies. This includes the development of AI-derived analytics, mobile intelligence, surveillance, and reconnaissance (ISR) cell development, and high-resolution synthetic aperture radar (SAR) imaging. Lockheed Martin will provide its Ground Antenna Transmit Receive (GATR) system for automated satellite imagery analysis, while Finnish firms will contribute their expertise on command and control, digital solutions, and radio frequency systems.</p>
 <p>The goal of this initiative is not just technological innovation, but also strengthening the security and aerospace capabilities of Finland and fostering partnerships with friendly nations and NATO Allies. Other top priority issues this initiative plans to handle are AI cybersecurity, with Lockheed Martin and the National Institute of Standards and Technology working on standards that will protect AI systems from hostile machine learning attacks.</p>
</div>

ICEYE partners with Lockheed Martin and Finnish firms to develop defense space technologies

AI, robotics, and IoT to streamline the flow of goods and information

Logistics Tech

<div>
 <p>Dematic, a warehouse automation solutions provider, has installed its AutoStore goods-to-person (GTP) system at South West Healthcare's Regional Logistics Distribution Centre in Warrnambool, Victoria. The implementation features selective pallet racking for large items and supports 13 hospitals in the region.</p>
 <p>The AutoStore GTP system uses robots to fetch totes from a storage grid and transport them to ergonomic pick stations, automating the picking process with order fulfillment capabilities. It sports a modular design to allow future growth. The system can accommodate up to 6,500 totes and requires 80% less space than conventional manual racking arrangements.</p>
 <p>With the new system, storage capacity has reportedly increased from one week to six to eight weeks for regularly used items, reduced picking errors due to automation, and improved order fulfillment times with same-day delivery for the main hospital and next-day delivery for regional branches. 24/7 remote IT support and quarterly maintenance are included in the system, provided by Dematic's service team based in Victoria.</p>
 
</div>

All Updates

Scale AI and OpenAI collaborate to fine-tune GPT-3.5 with custom data

Snowflake acquires Datavolo to enhance data integration capabilities for undisclosed sum

Snowflake acquires Datavolo to enhance data integration capabilities for undisclosed sum

OceanWell raises USD 11 million in Series A funding to build water farms

H launches Runner H, an AI agent for business automation

Capgemini partners with Mistral AI and Microsoft to expand GenAI solutions globally

Converge Bio launches biotech LLM platform with USD 5.5 million seed funding

Snowflake partners with Anthropic to integrate Claude AI models into Cortex AI platform

DeepSeek releases AI reasoning model DeepSeek-R1

ICEYE partners with Lockheed Martin and Finnish firms to develop defense space technologies

Dematic installs AutoStore system at South West Healthcare logistics hub

Machine Learning Infrastructure

Scale AI and OpenAI collaborate to fine-tune GPT-3.5 with custom data

Contact us