Cleanlab Receives $25M to Boost the Value of Every Data Point in Enterprise AI Solutions
Cleanlab is an automated data curation platform to enhance the utility of every data point in enterprise AI, LLM, and analytics solutions.
![Cleanlab Receives $25M to Boost the Value of Every Data Point in Enterprise AI Solutions](/content/images/size/w1200/2023/10/Cleanlab.jpg)
Company Name: Cleanlab
Location: San Francisco, CA
Product: Automated data curation platform to enhance the utility of every data point in enterprise AI, LLM, and analytics solutions.
Funding Details: $25M in Series A funding (Total funding to date: $30M)
Funding led by: Menlo Ventures and TQ Ventures; with participation from existing investor Bain Capital Ventures (BCV) and new investor Databricks Ventures.
Purpose of Funding: Expansion of operations and broadening of business reach.
Customers Include: AWS, JPMorgan Chase, Google, Oracle, and Walmart, as well as innovative startups like ByteDance, HuggingFace, and Databricks.
Leadership: Co-founded by Curtis Northcutt, Anish Athalye, and Jonas Mueller
About Company: Cleanlab has developed an automated data curation solution, Cleanlab Studio, which automatically adds smart metadata, substantially reducing manual labor and converting real-world data into practical model inputs.
By enhancing the dependability and profit margins of enterprise analytics, LLM, and AI decisions, Cleanlab plays a pivotal role.
One significant feature is the system's ability to automatically discern a majority of a dataset that's issue-free, thus elevating the profit margins of enterprise operations by dodging expensive data quality and annotations. Recently, they've introduced features targeting unreliable LLM outputs.
Their Trustworthy Language Model (TLM) delivers premium LLM outputs akin to those of ChatGPT, Falcon, and other similar LLMs, and also appends a trustworthiness reliability score to each output.
Cleanlab Studio's versatility has the capability to identify and rectify issues in various datasets like text, images, and tabular data.