News & Intel
LiveDaily AI-curated intelligence on Palantir Foundry, Ontology, AIP, Apollo, contracts, and community feedback. Updated automatically via GitHub Actions every day at 7 AM UTC.
Request for Data Engineer Associate Certification Exam Code
Hi, I recently passed the Palantir Foundry Aware Certification and would now like to take the Data Engineer Associate Certification. When I try to register, the certification portal requires a promo/exam code. I don’t have a Palantir point of contact. I also contacted the certification email address, but received an automated response saying that the inbox is unmonitored. Could someone please advise me on how I can obtain an exam code for the Data Engineer Associate Certification? Thank you. 1 post - 1 participant Read full topic
Frozen Speedrun AIP Workflow Deployment
I just started learning to use Foundry and am getting stuck at the Deploy step of Transforming the Data. Every time I deploy, even after I have deleted and rebuild the entire pipeline from scratch, my deployment gets stuck. I have tried it both with the data outputs grouped and ungrouped by changing settings in the Use LLM node. Has anyone encountered this before or have an idea of a workaround? I can’t find anything similar on any of the tutorials or youtube videos. . 1 post - 1 participant Read full topic
Faster Pipelines Out of Memory Errors - improve error message
My “Faster” Data Pipelines failed because of a OOM error. I feel like the error message in this cases could be improved. Also AIP didn’t detect correctly that it was a OOM error. Even though it was more or less visible in the logs. 1 post - 1 participant Read full topic
Minimum permissions for read-only OSDK access with an object security policy
I’m validating a browser OSDK/OAuth read-only test on Developer Tier. The object type is saved in a project and has an object security policy requiring the current user’s ID in an authorized-principal array. The tester has Hosted Website Viewer. Resource-level grants are disabled. What exact custom-role operations are required, granted only on this object type, for metadata retrieval, object lookup and filtered search—with no backing-dataset access, editing or resharing? Is ontology:view-object-type sufficient? Can foundry-data-proxy:get-files and security-rid-manager:write-secured-resources be omitted? Please also confirm the required OAuth read scope and how to verify whether separate download restrictions are enforced for my enrollment. The test must allow metadata retrieval while excluding another user’s case from lookup and search. 2 posts - 1 participant Read full topic
Early Talent Hiring at Palantir
What Hiring Managers value — and how they’ve built their careers at Palantir Editor’s Note: Technical Recruiter Rachel Vogel speaks with Hiring Managers to learn more about their career journeys and life at Palantir . Are you eager to join Palantir’s Early Talent program? Our Early Talent program focuses on engineering hiring for new grad roles and internships. In this blog post, we shine a spotlight on the amazing individuals who oversee the hiring process and give you a glimpse into their lives at Palantir. Get to know these accomplished professionals as they share their personal journeys, highlighting the mentors and experiences that have shaped their careers at Palantir. Learn how they embraced growth opportunities, became involved in the interview process, and forged meaningful connections with colleagues. Get ready to connect, learn, and feel empowered as you prepare to embark on your own transformative career at Palantir. Elliot K., Hiring Manager for Gotham Software E
Monitor Health Data in a Webapp
Hi, what is the best approach for displaying Foundry Dataset and Pipeline Health checks and monitoring-view alerts in a React web application? Do I need to set up web hooks and only monitor failures or can we use the internal endpoints that pages like the build page hits if we want to continuously monitor? 2 posts - 2 participants Read full topic
How to reslove "Failed to create connection to S3 bucket" error
I have been able to use this code to read files in media set in code repository in the public foundry: with control_file_input.get_media_item(item.mediaItemRid) as stream: xml_bytes = stream.read() In recent several days, the code will generate error message at the 2nd line: Failed to create connection to host host='pcloud-mio-REDACTED.s3.us-east-1.amazonaws.com', port=443 Can I adjust my settings in some way to solve this error? 1 post - 1 participant Read full topic
Unable to access Dev Tier in Mexico — Associate Application Developer exam code request
Hello, I recently passed the Palantir Foundry Aware Certification with a score of 93%, and I am now preparing for the Associate Application Developer Certification. I do not have a Palantir point of contact. I understand that AIP Developer Tier users can normally request an exam code by submitting an “Account Questions” Foundry issue through their Foundry instance. However, when I attempted to sign up for AIP Developer Tier, I received a message indicating that sign-up is currently not available in Mexico, so I am unable to access Foundry Support through that route. Could someone please advise me on the correct process for requesting an Associate Application Developer exam code or voucher in this situation? Thank you! 1 post - 1 participant Read full topic
ServiceNow vs. Palantir Technologies: Which Tech Stock Is a Better Buy in 2026?
As artificial intelligence continues to transform enterprise software, investors are choosing between the steady expansion of ServiceNow (NYSE:NOW) and the explosive growth of Palantir Technologies (NASDAQ:PLTR). This comparison explores which stock is the be…
Bug Report: Jupyter Notebook transform output registration UI is broken in Code Workspaces
Summary: Jupyter Notebook transform output registration UI is broken in Code Workspaces Steps to reproduce: Open a JupyterLab workspace in Code Workspaces In the Data panel, add a dataset intended as an output. Create a JupyterLab workbook, importing the dataset Attempt to register the workbook as a Report Create panel does not display an outputs section Instead outputs are listed with inputs The hotdog menu on the dataset row, selecting Mark as output — the dataset disappears from the panel entirely rather than moving to an outputs section. Expected behavior: The Edit Transform panel should display a clearly labelled outputs section, and marking a dataset as an output should move it there. Actual behavior: No outputs section is visible in the Edit Transform panel. Using “Mark as output” causes the dataset to vanish from the panel. Impact: Unable to register output datasets for Jupyter Notebook transforms through the UI, resulting in AliasNo
Oracle CDC RAC - Redo Thread is inconsistent
In Oracle CDC Streaming, data is collected normally for some time, then new events stop arriving with the following error: Redo Thread 1 is inconsistent; failed to find log with sequence 927401 On the Oracle side, Thread 1 / Sequence 927401 does exist and is currently available: ARCHIVED=YES DELETED=NO STATUS=A DEST_ID=1 LOG_ARCHIVE_DEST_1 is also VALID / LOCAL . This looks similar to Debezium issue DBZ-1801 related to archive log deduplication in Oracle RAC. Could you please confirm: Which Debezium Oracle Connector version Foundry is using Whether the DBZ-1801 fix is included or backported Whether this behavior could be related to the same RAC archive-log/thread handling issue For the same SEQUENCE# , records also exist for other redo threads, and some of them are DELETED=YES / STATUS=D . 1 post - 1 participant Read full topic
Feature Request: Custom Ordering of Properties in OMA
Overview I would like to propose an enhancement for Ontology Manager (OMA): introducing the ability to explicitly order Object Properties. The goal is to establish a deterministic, globally applied column and property order across all downstream Ontology applications in both developer and user modes (e.g., Object Explorer, Quiver/Insight, Workshop, and various property selectors). (See also my previous request in 2024: https://community.palantir.com/t/set-default-columns-order-in-ontology-editor/1974) Current Behavior Currently, there is a lack of synchronization regarding how properties are ordered across the platform. The column order in the backing dataset, the properties displayed within OMA, the default column order in Insight tables, and the layout in property selectors often appear in completely different sequences. (See the attached picture for an example of this misalignment). Proposed Solution Allow ontology builders to define and lock a primary, default
Feature Request: Display Backing Property Name for the "Title" Special Property
Overview I would like to propose an enhancement for applications that display Ontology Object Types (such as Ontology Manager, Object Explorer, Workshop, etc.). The request is to append the underlying property/column name next to the generic “Title” label for the designated special title property. Current Behavior Currently, when looking at an Object Type’s properties in various applications, the property designated as the primary identifier is simply displayed as "Title" . There is no immediate visual indication of which underlying column or property was actually used to populate it. Proposed Solution Update the UI to include the name of the backing property in parentheses next to the special Title label. For example, if the Tool ID column is configured as the Title property in the Ontology Manager, the display should read: "Title (Tool ID)" instead of just "Title" . Key Benefits & Business Value User Transparency & Context: End-users often need to k
Bug Report: Dates in Pipeline Builder
I’m working on an SAP migration that uses the placeholder date Sep-30-9999 (i.e. FY9999) for undefined expiration dates. When I enter this in Pipeline Builder, the picker displays Sep-30-1826. 1 post - 1 participant Read full topic
Global branch visibility across platform
Hey all, Ive got an opportunity to play around with global branching recently. It really is a game changer compared to before. But there are few drawbacks that I came across(or am I missing a setting/maybe the feature is not available in my stack just yet) - Post transform is build in global branch, the dataset can only appears as a rawdataset in files as it is not directly possible to move around folders/files using global branch selected (I assume its “Main” by default as there is no option to make a selection) The global branch in Ontology Manager did not make the newly created dataset available for selection when trying to create an object type (as it was not discoverable to pick using the “use existing datasource → Select datasource” which is the usual flow to create an object type) Another drawback is when working in “Workshop” using the global branch; it does not automatically save changes On the flip side, the best part is seeing how modified, updated,
Model Studio - Out of Memory during training & branches support
Hello, I have prepared a dataset for use with Model Studio for a time-series prediction. My input data is around 3GB - 19M rows however I keep getting OoM issues even when memory is bumped to 64GB + A10G GPU. While I understand that I could reduce the size of my input data, there is no data loader implemented ? As well, it would be great to support dataset branch as input. Best regards, 1 post - 1 participant Read full topic
SuperRepo deploy: imported types become install inputs that don't validate (SPT references, many-to-many link sides)
Follow-up to our local-preview post. We tried to deploy the same SuperRepo (three code-born types, two function-backed actions, three TypeScript v2 functions, a React app) that imports ~70 object types read-only from an existing Ontology Manager ontology. The imports are mandatory install inputs, which makes sense — but we could not get them to validate in an install draft. We ran draft-only checks (upload signed bundle → install draft → read validations → delete draft) and have parked the SuperRepo route for this app for now. Three findings, in case they help: **Environment:** foundry-cli 0.253.0 · ` @osdk /maker` 0.67.0 · ` @osdk /maker-experimental` 0.61.0 · ` @osdk /maker-import` 0.37.0 · lock written by `foundry import ontology --objects …`. **1. Imported properties reference shared property types the bundle never declares.** Every imported property backed by an SPT gets `“sharedPropertyType”: “”` in its input shape (135 references in our bundle), but no bl
How to obtain a "trace" of an AIP Logic, to display in Workshop?
I want to see a “chain of thought” of an AIP Logic in Workshop. I want to see the steps some LLMs blocks took, which was the exact steps of the execution of the function, what was decided and why. How can I capture the execution steps of an AIP Logic ? How can I display this in Workshop ? 2 posts - 1 participant Read full topic