This page is for the participants of the (Sponsor name and challenge) hackathon taking place at (Event) - please use this space for all information and further training. 

Unlock agentic-era autonomous networking with Google Cloud AI and TM Forum code and frameworks


Your mission is to build modular, intelligent network-related prototypes addressing real-world CSP challenges. Teams will combine standardized TM Forum code frameworks (like Agents, Intent, Closed Loop, Open APIs),  AN Assets (AN Blueprint, DT4DI toolkits etc.), with Google Cloud AI and AI optimized infrastructure. 

The goal is to build modular, intelligent network-related use cases that demonstrate progress towards autonomous network level 4; focusing on enhancing network automation, sustainability, or customer experience, leveraging technologies like Gemini 2.5 and agentic systems. 

Teams need to deliver a working prototype/demo and a short pitch explaining the problem, solution, impact, and architecture.

Teams should aim to deliver and showcase:

  • Working prototype or demo
    (e.g., GenAI Network Operations Centre assistant, root cause analysis engine, anomaly detection, sustainability dashboard, etc.)
  • Use of TM Forum Standards & APIs
    Demonstrate alignment with TM Forum’s Open Digital Architecture and Autonomous Network L4 capabilities.
  • Google Cloud integration
    Integration with Google Cloud Platform (GCP) services for collaborative agents, Data, AI/ML..

Short pitch/demo
Clearly explain the problem solved, real-world impact, architecture,  and a roadmap to production deployment.


Sponsored by 



1. Hackathon Timeline/Overview 

Key Dates

  • May 30: Registration deadline
  • Upon Registration: Briefing deck supplied, pre-learning, choose your challenge, environment & data prep, initial meeting with Google mentor
  • June 4 - Webinar, overview and onboarding
  • June 2 - 16: Continue preparations; email communications and weekly meetings with the mentor
  • June 17 - 19: Prototype Development: 9am – 5pm, Bella Center, Copenhagen
  • June 19 : The Pitch, 9am – Midday Awards – The Competition Winner is announced  12:30pm


July - September

Demos showcased on TM Forum website 

Social media exposure

Spotlight on TM Forum Inform for the winning team

Webinar with winning team hosted by TM Forum and Google 

2. Use case inspiration

N.B - Use case ideas below are just illustrative, please bring your own ideas to address you sepcific challenge

Predictive Fault Prevention & Automated RCA

    • Problem: Reactive RCA is slow, downtime costly.
    • Hack:  Predict failures via telemetry, automate RCA correlation.
    • Autonomous Actions: Trigger diagnostics, isolate component, self-heal, create ticket

AI Config Gen & Digital Twin Validation

    • Problem: Manual config is error-prone, slow validation.
    • Hack: Translate intent to config, validate in digital twin.
    • Autonomous Actions: Generate config, test in twin, report validation.


.


Dynamic RAN Traffic Steering & Energy Saving

    • Problem: Static RAN capacity wastes energy.
    • Hack: Monitor load/patterns, dynamically adjust RAN parameters.
    • Autonomous Actions: Calculate optimal parameters, issue RAN commands.


Check out TM Forum’s latest introductory guide to Autonomous Network L4 High Value Scenarios v2.0 for more information

3. Recording of the Onbaording webinar 

For those who could not attend the onboarding webinar please see the recording here 

4th June Webinar recording.mov


4.  Mentors and TM Forum contacts


Team

Google Mentors

TM Forum Contacts

Orange

Antoine Blanvillain / Laurent Juilliard

Olta Vangjeli

Dmytro Gassanov

Telefonica

Yolanda Azcunaga / Angel Garcia

Olta Vangjeli

Dmytro Gassanov

Telenor

Juan Ramirez / Hans-Christian Fuchs

Olta Vangjeli

Dmytro Gassanov

Telus

Abed Ibrahim / Glen McGowan

Olta Vangjeli

Dmytro Gassanov


5. Technical Training - Qwiklabs Course

Qwiklabs course

Course Title: Google Cloud DTW '25 Hackathon: Practice Labs

Course Description:

Welcome to the Google Cloud DTW '25 Hackathon Training Preparation Course!

This essential (and optional) pre-course on Qwiklabs is meticulously designed to equip you with the fundamental technical knowledge and hands-on experience necessary to excel in the upcoming Google Cloud DTW '25 Hackathon event.

This course contains carefully curated links to labs and other valuable resources that will help you prepare for the core technical topics and Google Cloud services you'll encounter during the hackathon. It's broken down into sections and topics that directly align with the major tasks and challenges your team will be expected to complete during the event, specifically leveraging Google Cloud technology enablers for autonomous networking solutions.

Why take this course?

  • Build Foundational Skills: Gain practical experience with key Google Cloud technologies (like AI/ML services, data platforms, networking tools) for the hackathon's objectives.
  • Validate Your Knowledge: Earn skill badges by completing hands-on labs and challenge assessments, validating your proficiency in specific Google Cloud products and solutions.
  • Boost Your Confidence: Feel fully prepared and confident to tackle the hackathon challenges from day one, knowing you have a solid technical foundation.
  • Enhance Your Profile: Badges prove your professional capabilities and can lead to increased career opportunities within the Google Cloud ecosystem.

 

How to get the most out of this course:

Work through the labs at your own pace. Each lab is designed to provide hands-on experience and reinforce the concepts necessary for success in the Hackathon. You can earn a badge by completing a course, or for those already familiar with the concepts, you can jump straight into the challenge lab to get your badge today!

Get ready to innovate with Google Cloud at DTW '25!

Steps to access the optional pre-work labs:

  1. Go to explore.qwiklabs.com
  2. Create an account.
  3. Browse togoo.gle/dtw2025-hackathon to access the pre-work labs

 

Qwiklabs FAQs for the Hackathon

Q1: What is Qwiklabs and how does it relate to this hackathon?

A1: Qwiklabs provides hands-on labs that allow you to work directly within real Google Cloud environments. For this hackathon, we're offering free pre-hackathon Qwiklabs access specifically designed to help you gain practical experience with Google Cloud technologies before the competition officially begins. It's a great way to learn, brush up on skills, and explore services relevant to your project.


Q2: How do I access my free Qwiklabs account for the hackathon?

A2: Instructions on how to redeem your free Qwiklabs access will be shared by email. Please refer to that email instructions for the specific steps and a unique enrollment link.


Q3: What kind of labs are available on Qwiklabs?

A3: You'll find labs on technologies like ADK, Spanner

Q4: Is completing Qwiklabs labs mandatory for the hackathon?

A4: No, completing Qwiklabs labs is optional. However, we highly recommend taking advantage of this free resource. It's an excellent opportunity to build confidence, explore new services, and gain practical skills that will give you an edge during the hackathon.


Q5: How does Qwiklabs differ from the Google Cloud Project with $2,000 credits, or the team's "scratchpad" project?

A5: It's important to understand the distinctions:


Qwiklabs Environments: These are separate, temporary environments provided for learning purposes only before the hackathon. They come with pre-configured scenarios and are ideal for following guided labs.


Team "Scratchpad" GCP Project: This is a dedicated GCP project for your team to use for early experimentation, testing, and initial prototyping during the hackathon, before committing to your final design.


Main GCP Project with $2,000 Credits: This is your  primary environment (that you’ll bring to the hackathon),  utilizing the $2,000 in Google Cloud credits for the competition.


Qwiklabs is primarily for pre-hackathon learning, while the scratchpad and credit-based projects are for development during the hackathon.


Q6: Can I use Qwiklabs during the actual hackathon event?

A6: Yes, qwiklabs environment will be available up until June 27th, 2025


Q7: What if I encounter issues or have questions while using Qwiklabs?

A7: Qwiklabs typically has its own in-platform support or help resources. For general queries about accessing Qwiklabs, please refer to the instructions on the hackathon Confluence page. If you have specific issues with a lab environment, look for the 'Need Help?' or 'Support' options within the Qwiklabs platform.


Q8: Will completing Qwiklabs labs earn me "Skill Badges" that I can display?

A8: Yes, if you complete certain "Quests" (learning paths) on Qwiklabs (now part of Google Cloud Skills Boost), you can earn Skill Badges which validate your hands-on experience and can be shared on your professional profiles like LinkedIn. While not required for the hackathon, it's a great way to showcase your newfound skills!

6. Techinal Training and resources

 Start Self Learning & Tech Stack - Google Cloud

Resource  ITEM  Description 
1. Agentic Architecture & Models Enable autonomous behaviour using collaborative, goal-driven agents and powerful foundation models.
2. Data, Telemetry & Topology
  • Spanner Graph Platform: Used to build digital twins and represent complex relationships across network elements, users, and incidents.
  • Dataflow: Real-time data processing of streaming telemetry, events, and KPIs.
BigQuery: Analytical engine to query historical patterns and performance data at scale.
Power data-driven intelligence across networks using scalable, real-time pipelines and relationship-aware models.
3. Visualisation Layer
  • Looker / Dashboards: Interactive NOC-style dashboards to visualise service status, anomalies, agent decisions, and digital twin dynamics.
Spanner Graph Visualisation: Graph-based UI for exploring entity relationships and topological dependencies.
Provide NOC operators and agents with actionable visual insights into network behaviour and digital twins.
4. Deployment Environment
  • Serverless: Google Cloud allows CSPs to not think about creating, scaling, securing and managing agents, AI/ML, data and other services using Agent Engine, Vertex AI and so on, to focus on the business problems.
  • Cloud Run: Lightweight hosting for micro-services supporting twin-to-agent interfaces and domain-specific logic.
  • GCP Canvas / Workbench (optional addition): A visual platform for orchestrating pipelines, agents, and services collaboratively.
Flexible runtime environments to scale services and integrate modular twin/agent capabilities.
5. Google Proof-of-Value RAN Agent 

We are excited about offering an agent that we have built for you to optionally use to enhance your hackathon scenario.  I have attached a PDF user guide for you to deploy and start test driving the agent.  

We've built a powerful Proof-of-Value (PoV) RAN agent designed to revolutionize how you approach telecom use cases, and we invite you to put it to the test at this hackathon!

What is it?
Our intelligent agent understands complex RAN performance indicators and comes equipped with a sophisticated anomaly detector. Leveraging in-house machine learning advancements from Google and the power of Large Language Models (LLMs), our agent can help you build and solve telecommunications challenges with unprecedented speed. Think of it as a fast-track to identifying and diagnosing critical network incidents.
Key Capabilities:
  • Understands RAN KPIs: No need to be a deep domain expert. Our agent can interpret and reason about RAN performance counters.
  • Built-in Anomaly Detection: Quickly surface anomalies related to common RAN issues.
  • Pre-trained on Use Cases: We've trained the agent on identifying issues like RAN congestion, capacity limitations, and call failures, enabling rapid incident identification.
Your Data, Your Use Case:
While our agent has been trained and tested on a foundational set of RAN performance management counters, the real magic happens when you bring your own data. As part of this hackathon, we invite you to provide your own relevant datasets to see how the agent performs on your specific use cases. The more data you bring, the smarter it gets!
You're in Control:
This isn't a black box. Our RAN agent will run directly in your tenant, giving you complete control over its behavior and the data it accesses.

Come and see how our PoV RAN agent can give you a winning edge. Let's build the future of network operations together! 



 Start Self Learning & Tech Stack - TM Forum


TM FORUM Autonomous networks Powerpoint Presentation - AN Google Hackathon TMF assets.pptx

Resource  ITEM  Description 
1. Open APIs 

TM Forum - Open API Directory

  • TMF 921 - Intent Management API
  • TMF 753 - Closed Loop Management
  • TMF 641 - Service Ordering Management API
  • TMF640 - Service Activation Management API
    ............ 

Standardized interfaces for Intent, service inventory, product catalog, service activation, resource inventory, etc.

Enables seamless integration between mock CSP systems and GCP-hosted agent/twin logic. Ensures modularity and composability.

2. AN Architecture 

TM Forum Autonomous Networks

AN Architecture

AN Agent Architecture


Observability and Monitoring

Autonomous Networks - a set of network and software platforms that can sense its environment and
adapting its behavior accordingly with little or no human input.

Autonomous Domain - serves as the basic unit that can fulfill closed-loop automation of specific
network operations.

3. AN Use Cases and Blueprints  

Autonomous Networks: Level 4 industry blueprint – High Value Scenarios

IG1339 Autonomous Networks L4 High Value Scenarios 

IG1373 AN Use Cases: A Guide to Self-Healing and Closed-Loop Automation v1.3.0

IG1369 GenAI Use Cases v1.3.0

Defines requirements for L4 capabilities (intent, self-X, closed loops, agent frameworks).

Use as a reference architecture to validate whether the hackathon prototype reaches L4 autonomy milestones by implementing the suggested design pattern and enablers. 

4. AN Intent 

TM Forum - Intent-based Automation  

Intent User Guide and Introductory Guides:

Turtle Files Collection TIO v3.6.0

  • Intent Ontology
  • Intent Common Model
  • Intent Extension models
  • Intent Specification

Structured representation of goals, intents, and policies in business, service and resource layer.

Supports the design of GenAI interfaces or agents that translate human/network goals into machine-executable actions. Encourages intent-driven systems

5. Digital Twin for Decision Intelligence Use Cases

IG1310C - DT4DI Top Use Cases v10.0.0

  • DT4DI Use Case Classification
  • DT4DI for StrategyDT4DI for Marketing and Sales
  • DT4DI for CEM & Customer Management
  • DT4DI for Revenue Management
  • DT4DI for Service Management
  • DT4DI for Business Partner Management and B2B2X
  • DT4DI for Infrastructure Management
  • DT4DI for Enterprise Management

IG1307 Digital Twin for Decision Intelligence (DT4DI) Whitepaper v2.0.0

Toolkit for creating cognitive digital twins with real-time inference and decision-making abilities.

Supports modeling of the network state and integration of ML/AI agents to optimize operations, e.g., RCA or sustainability KPIs.

7. On site Logistics (Agenda, Location, Internet etc) 

Location onsite and access - 

The hackahton will take place in the hackaton quad which can be found in hall C4 of the bella center here on the floorplan https://dtw25-ignite.expofp.com/?hackathon
Please note not all of the team is required on site and a hybrid set up can be used, we would suggest a minimum of 2 people on site 

EVENT QUAD OPENINIG TIMES & SCHEDULE 

  • Tuesday, 17 June: Quad open 8:30 am – 7:00 pm

  • Wednesday, 18 June:  8:30 am – 6:30 pm

  • Thursday, 19 June: 9:00 am – 2:30 pm

Internet Onsite - 

LAN wired connections will be provided as a standard and we strongly recommend you use the wired connections whenever possible, as the Wi-Fi may have disruptions beyond our control. 

laptops and many mobile devices that do not have a built in ethernet ports can use USB-C adapters (for example) and you should bring these with you to use the wired connections.

Agenda - 

TIME DAY 1 "Prep & Hack" Day 2 "Hack Continues" Day 3 "Reward"
AM  09:15 Welcome Briefing  Hacking ....(avaiable mentor support) Hacking STOP
AM Participant team ideation and 
brain storming 
Hacking .... Hackathon pitch to judges 
AM Hacking starts Hacking ....(avaiable mentor support) Scored collected 
PM Hacking ....    Hacking ....
PM Hacking .... (avaiable mentor support) Prepare elevator pitch & Demo

Award ceremony and winner announced 

PM Hacking ....  Hacking .... RELAX


8. Judging (Agenda/ Criteria / Submission deck) 

Judging will take place on the 19th June at 09:00 – 11:00 onsite in Hall B , meeting room M4 which can be found here on the floorplan 

The team will be presenting their use case and demo in a 10-minute pitch with a 5-minute Q&A session,  SUBMISSION TEMPLATE HERE 

  1. Use case concept & Architecture –A brief description of what the use case is and a simple view on the architecture implemented to deliver the use case. 

  2. Business Impact Analysis - How much impact would this use case bring if deployed at scale?

  3.  Use case demo – A recording of the work in action

  Judging agenda & Team time slots for the 19th June at 09:00 – 11:00

Time

Agenda

09:00 - 09:10

Opening remarks 

10min Presentation - 5min Q&A

Team one 

10min Presentation - 5min Q&A

Team two 

10min Presentation - 5min Q&A

Team three 

10min Presentation - 5min Q&A

Team four 

10min Presentation - 5min Q&A

Team five

10:30 Close of judging session


10:30 - 11:00 Judges closed scoring session 

Closed judges scoring session and scores collected 

 

.

Judging Criteria 

Winners will receive:

  • 3 months of follow up support from a Google mentor

  • $6,000 USD GCP credits

  • Trophy + Google Cloud Swag

  • TM Forum Inform article spotlight

  • Post event marketing - ‘winning social posts

9. Submission Folder 

At the end of the Hackathon please deposit your presentaiton and relevant documentation in the folder here SUBMISSION FOLDER  


10. How to redeem your $2,000 Credits 

COUPON REDEMPTION: To redeem promotion code, customers must go to console.cloud.google.com/billing/redeem.  

Coupon codes are redeemable by both new users of GCP and existing customers. New customers will need to create a BILLING ACCOUNT in order to redeem coupons (which can be done at console.cloud.google.com/billing/redeem). Existing customers should have an existing BILLING ACCOUNT. A maximum of ONE coupon code per billing account can be applied for coupon codes from the same program.

Customers can raise any issues with billing / coupon redemption directly with the Billing team at https://cloud.google.com/support/billing


Please note, the following SKUs are excluded from marketing coupon and credit offerings: https://cloud.google.com/skus/exclusions. In particular, Gemini API on Google AI Studio is NOT covered under these coupons. This means the following SKUs are excluded from the promotion:

Generate_content_output_token_count_gemini_1.5_flash_short_input

Generate_content_output_token_count_gemini_1.5_pro

Generate_content_input_token_count_gemini_1.5_flash_short_input

Generate_content_cached_input_token_count_gemini_1.5_pro_short_input

Generate_content_cached_content_storage_token_hours_gemini_1.5_flash

Generate_content_cached_content_storage_token_hours_gemini_1.5_pro

Generate_content_input_token_count_gemini_1.5_flash

Generate_content_input_token_count_gemini_pro

Generate_content_input_char_count_gemini_pro

Generate_content_cached_input_token_count_gemini_1.5_flash_long_input

Generate_content_input_token_count_gemini_1.5_flash_long_input

Generate_content_output_token_count_gemini_1.5_flash_long_input

Generate_content_output_token_count_gemini_pro

Generate_content_input_token_count_gemini_1.5_pro_long_input

Generate_content_input_token_count_gemini_1.5_pro_short_input

Generate_content_output_token_count_gemini_1.5_flash

Generate_content_cached_input_token_count_gemini_1.5_flash_short_input

Generate_content_output_token_count_gemini_1.5_pro_long_input

Generate_content_output_token_count_gemini_1.5_pro_short_input

Generate_content_cached_input_token_count_gemini_1.5_pro_long_input

Generate_content_output_char_count_gemini_pro

Generate_content_input_token_count_gemini_1.5_pro

  • No labels