Issue - https://issues.apache.org/jira/browse/RANGER-5278 has been created with 
high level requirements.
I will create a separate thread to discuss the implementation details and 
refinement of requirements  if needed.

Thanks,
Selva-


> On Jul 29, 2025, at 4:50 AM, Selvamohan Neethiraj <[email protected]> wrote:
> 
> Hi,
> 
> I am currently working on submitting the initial Jira ticket with the 
> requirements and a high-level plan for the next steps as part of our 
> discussion with the ranger community. The ticket should be created by the 
> first week of August, so that we can schedule further discussions.
> 
> Regards,
> Selva-
> 
>> On Jul 29, 2025, at 1:38 AM, Fateh Singh <[email protected]> wrote:
>> 
>> Hi Selva
>> Do you have any next steps or a plan in mind to execute on these AI
>> features? Any JIRAs you have filed ?
>> 
>> Regards
>> Fateh Singh
>> 
>> On Thu, Jul 17, 2025 at 10:49 PM Kumar Basapuram <[email protected]>
>> wrote:
>> 
>>> Hello Selva,
>>> 
>>> Looking forward to contribute and learn from it.
>>> Thanks
>>> 
>>> 
>>> 
>>> *Thanks & Regards*
>>> 
>>> *Kumar Basapuram,*
>>> *Mobile: +91-9962904416*
>>> *Mail-to: [email protected] <[email protected]>*
>>> *Banglore -India.*
>>> 
>>> 
>>> On Fri, 18 Jul 2025 at 11:08, Fateh Singh <[email protected]> wrote:
>>> 
>>>> Hi Selva
>>>> 
>>>> +1 for AI initiatives and looking forward to contributing to these
>>>> initiatives.
>>>> 
>>>> Here are a few more use cases we can consider-
>>>> 
>>>> 1) Explainable Ranger Authorization i.e why ranger is making a certain
>>>> authorization decision - Other Authorization products such as OsoHQ are
>>>> already advertising this as a novel feature of their product
>>>> https://www.osohq.com/post/how-we-built-an-interactive-debugger. I have
>>>> some ideas on how we can use LLMs to achieve this for Ranger.
>>>> This should also help customers understand if a user will have access to
>>> a
>>>> certain piece of data based of the current ranger admin data
>>>> 
>>>> 2) AI ranger Admin: Natural language/Prompt based policy creation - User
>>>> gives their use case based on which AI should be able to create
>>> policies. A
>>>> simple example: define a policy in a way that a user cannot access the
>>>> salary and gender data at the same time
>>>> 
>>>> 3) Not a user facing feature but for developers only - in my experience,
>>>> LLMs in their current form right now are not good at assisting developers
>>>> for Ranger development atleast for slightly more complex use cases. How
>>> and
>>>> what knowledge can we give to LLMs so that it can be better suited for
>>>> aiding the developers to develop new features. This initiative will
>>> include
>>>> prompt engineering, knowledge base creation and chat with Ranger
>>>> documentation.
>>>> 
>>>> Would love to discuss what can be the next steps for these initiatives.
>>>> 
>>>> Regards
>>>> Fateh
>>>> 
>>>> 
>>>> On Thu, Jul 17, 2025 at 10:32 AM Dilli Dorai <[email protected]>
>>>> wrote:
>>>> 
>>>>> I look forward to collaborate and contribute.
>>>>> 
>>>>> On Thu, Jul 17, 2025 at 10:30 AM Dilli Dorai <[email protected]>
>>>>> wrote:
>>>>> 
>>>>>> Hi Selva,
>>>>>> +1.
>>>>>> Great, timely initiative.
>>>>>> Regards
>>>>>> Dilli
>>>>>> 
>>>>>> On Thu, Jul 17, 2025 at 10:12 AM Ramesh Mani <[email protected]>
>>> wrote:
>>>>>> 
>>>>>>> +1 for Building agentic AI for Apache Ranger.
>>>>>>> 
>>>>>>> Thanks Selva for this initiative and discussion. It will be a good
>>>> value
>>>>>>> addition to the Apache Ranger project. Look forward to collaborating
>>>> and
>>>>>>> contributing.
>>>>>>> 
>>>>>>> Thank you.
>>>>>>> Ramesh
>>>>>>> 
>>>>>>> On Thu, Jul 17, 2025 at 8:17 AM Selvamohan Neethiraj <
>>>>> [email protected]
>>>>>>>> 
>>>>>>> wrote:
>>>>>>> 
>>>>>>>> All:
>>>>>>>> 
>>>>>>>> Hello,
>>>>>>>> 
>>>>>>>> I am initiating a new initiative to construct an Agentic AI for
>>>> Apache
>>>>>>>> Ranger, which will proactively monitor and enhance the Apache
>>> Ranger
>>>>>>>> Protections for the enterprise data.
>>>>>>>> 
>>>>>>>> Some of the key concepts associated with this initiative  are:
>>>>>>>> * auto-create of ranger policies based on access logs and
>>> obtaining
>>>>>>>> approval from the resource owner
>>>>>>>> * Analyzing the behaviors of users and roles and providing
>>>>> corresponding
>>>>>>>> reports
>>>>>>>> * Implementing real-time anomaly detection
>>>>>>>> 
>>>>>>>> Please provide your feedback. If we receive sufficient community
>>>>>>> support,
>>>>>>>> we can proceed with a preliminary proof of concept to assess the
>>>>>>>> feasibility of elevating Apache Ranger to the next level.
>>>>>>>> 
>>>>>>>> Thank you,
>>>>>>>> Selva-
>>>>>>>> Apache Ranger PMC.
>>>>>>>> 
>>>>>>>> 
>>>>>>> 
>>>>>> 
>>>>> 
>>>> 
>>> 
> 

Reply via email to