A flagship event of Bengaluru Tech Week 2026
Buildathon 2026
Powered by
Build something that survives a change of plan.
One day. A working product. And a non-negotiable new constraint at exactly 12:00 noon.
Choose one of Entire’s three open-source challenges and turn an idea into a working developer product. Build the first version, face The Noon Curveball and show how your team adapts without losing what already works.
Teams may also use Databricks within the same project and compete for an additional ₹50,000 in Best Use of Databricks awards.
Sunday, 6 September 20269:00 AM–5:00 PM ISTEntire main challenge: ₹2,00,000 in Amazon vouchersBest Use of Databricks: additional ₹50,000 in Amazon vouchers
Venue
Scaler School of Technology
RMQ9+HRM, 2nd Cross Rd, Electronic City Phase I, Electronic City, Hebbagodi, Karnataka 560100, India
Open in Google Maps ↗You’re in good company
Developer platform partner
Data + AI platform partner
Innovation partners
Platform partners
An official event of
Community partners
Communities running the floor with us, sending mentors, and bringing their people along for the last day.
The week talks. The last day builds.
Technology events are full of ambitious ideas. The Buildathon asks what happens when those ideas have to survive contact with a user, a deadline and a working product.
- 01Who experiences the problem?
- 02What is difficult or inefficient today?
- 03What should become possible through the proposed solution?
- 04What is the smallest useful experience that can be built and tested in one day?
- 05What evidence would show that the idea deserves to continue after the event?
The result does not need to be a production-ready company or a feature-complete application. It should, however, be more than a concept deck. A strong Buildathon submission should contain a working core experience, a clear explanation of the problem and transparent documentation of what was created during the event.
You have seven hours from the official start to code freeze, including setup, testing and submission. Choose a specific user, find a problem worth solving and build the smallest useful version. At noon, the brief changes. Your team must revisit an assumption, adapt the implementation and prove that the result still works.
The final hour is for demonstrations and judging. Show us the product, the decisions behind it and the evidence that your adaptation holds up.
What makes it different
A real change of requirements
The Noon Curveball becomes part of your challenge, not an optional bonus.
A visible development journey
Preserve your morning’s decisions so a fresh agent session can continue from them.
A working result
Demonstrate one complete user journey and test its important behaviour.
A reason to keep building
Explain what remains unresolved and what it would take to make the project useful beyond the day.
Who it's for
Team size, participant eligibility and the selection process are set by the participation rules linked from the application. The event is intended for people who enjoy turning ideas into tangible outcomes, including:
- 01Software engineers and AI builders
- 02Product managers and product-minded developers
- 03Designers who can shape usable product experiences
- 04Data practitioners and technical researchers
- 05Students with practical building experience
- 06Founders and independent makers
- 07Domain experts who can define meaningful real-world problems
Participants should be comfortable contributing to a team under a limited timeframe. Teams do not need every possible specialisation, but they should be able to divide responsibilities, make decisions quickly and deliver a demonstrable outcome.
Read the participation rules linked from the application before submitting. They explain team size, participant eligibility, selection and the application timeline.
Choose your challenge
Three Entire tracks. One optional Databricks award. One noon reveal.
Choose one of Entire’s three original open-source challenges. Build one project and submit it under that track. Every project must satisfy Entire’s technical and submission requirements.
Teams may also use Databricks for a meaningful part of the same product and opt into the Best Use of Databricks awards. Databricks is an additional award category across all three tracks, not a separate track or competition.
Build Entire: Open Source for the Agent Era
AI coding agents can produce code quickly. Developers still need to understand why it changed, review its impact and continue the work with confidence.
Build an original feature, integration or developer experience on top of the Entire open-source ecosystem. Your product must make Entire essential to what it does. Existing issues can inspire your idea, but you do not have to pick one.
- E1
Build a Checkpoint-Native Developer Experience
Build on: Entire CLI
Turn the context captured in real Entire Checkpoints into a useful developer workflow. Help someone review an implementation against its intent, identify unfinished requirements, assess a change or hand work to another developer or agent.
Your product must use actual Checkpoint context as an essential input. Simply using Entire to track development of an unrelated app does not qualify.
- E2
Build with Graph Intelligence
Build on: Entire Graph
Use the structure of a codebase to support a useful decision or action. Possible directions include impact-aware review, test selection, migration planning, codebase onboarding or an experience that connects Graph findings with Checkpoint intent.
You can improve Graph itself or build on its commands and output. Displaying raw output alone is not enough: turn the evidence into a workflow people can use and verify.
- E3
Bring Entire to a New Agent or Workflow
Build on: Entire External Agents
Connect Entire to a coding agent, editor, development environment, CI process or software-delivery workflow. Capture or use meaningful development context: prompts, responses, file changes, tool activity, usage information or lifecycle events.
External SDKs and fixtures are allowed, but your submitted implementation must live in the designated Entire fork. A thin interface that only invokes an Entire command does not qualify.
- Begin implementation only after the official start, in a new fork of your selected Entire repository.
- Mirror that fork on Entire, select the India region and work in the clone created from the mirror.
- Enable Entire Checkpoints before agent-assisted development. Preserve the initial plan, pre-noon stable state, curveball response and final verification.
- Use Entire Graph for investigation, impact analysis before risky changes and a final semantic-diff review. Verify its conclusions against source code and tests.
- Deliver a working end-to-end path, meaningful tests, a reproducible setup and an honest account of limitations.
- Complete The Noon Curveball in a fresh agent session using earlier Checkpoint context and Graph analysis.
Whichever track you choose, every submission should aim to demonstrate:
- 01A clearly defined user or stakeholderThe team should know who the product is for and what that person needs.
- 02A meaningful problemThe project should address a real difficulty, opportunity or unmet need rather than adding technology without purpose.
- 03A focused solutionThe team should choose a scope that can produce a credible result within the event window.
- 04A working core experienceAnother person should be able to interact with, inspect or meaningfully evaluate the central product idea.
- 05Transparent implementationThe submission should identify prior work, templates, external services, AI-generated components and open-source dependencies.
- 06A practical continuation pathThe team should understand what must happen next if the idea is to become a reliable product.
Best Use of Databricks
One optional award across all three Entire tracks.
Every team participating in one of the three Entire tracks may optionally compete for the Best Use of Databricks awards. Teams build one project and submit it once under their selected Entire track.
To qualify, Databricks must power a meaningful part of the product’s working data or AI workflow. Removing the Databricks implementation should materially reduce the product’s functionality, reliability or usefulness.
A team can win both an Entire main challenge prize and a Best Use of Databricks prize. Using Databricks is optional and does not change the requirements or judging criteria of the selected Entire track.
Eligible Databricks entries should demonstrate
- A working Databricks implementation connected to the submitted product.
- A clear explanation of which Databricks capabilities were used and why.
- Permitted public, synthetic or team-owned data with sources and limitations documented.
- A reproducible demo or deployment that judges can inspect.
- How the Databricks implementation responded to The Noon Curveball.
- The team’s own understanding of the architecture, code and trade-offs.
What does not qualify by itself
- Mentioning Databricks in a presentation without a working integration.
- Uploading data that the product does not meaningfully use.
- Using Databricks only as superficial storage or hosting.
- Showing a mock-up of an integration that judges cannot inspect.
- Adding an unrelated component solely to enter the award category.
Best Use of Databricks is an additional award category, not a fourth track. The same project is assessed for both awards, and Databricks use does not affect the Entire score.
At 12:00 noon, the brief changes.
Every team receives an additional, track-specific constraint at exactly 12:00 noon IST on 6 September. The exact requirement stays under wraps until the reveal.
It will challenge an assumption in your current design. Your job is to adapt the project you have already built, preserve its useful behaviour and demonstrate that the new requirement works.
- Before noon
Commit your last stable version and preserve enough context for a fresh session to understand the intent, architecture, completed work and open risks. Confirm that the commit has a valid Entire Checkpoint.
- At noon
Stop implementation, close the current agent session and receive your constraint. Start a fresh session, reconstruct the earlier work and assess the impact before editing. Use your Checkpoint and Entire Graph for this step.
- After noon
Implement, test and explain the adaptation. Show what changed, what stayed intact and why the result can be trusted.
The constraint is mandatory for every team and contributes to the judging score.
The build day
Sunday, 6 September • All times IST. Arrive ready to begin at 9:00 AM. Venue-entry and check-in instructions will be included with your participation confirmation.
- 8:00–9:00 AM
Arrival, breakfast and check-in
Doors open. Get set up before the official start.
- 9:00–9:30 AM
Official start, scope and setup
Teams fork, mirror, clone and verify their tools.
- 9:30–11:45 AM
Build the first working version
Find the smallest complete user journey and make it work end to end.
- 11:45 AM–12:00 noon
Commit the stable state
Preserve the handoff context and verify required Checkpoints. Confirm that the commit has a valid Entire Checkpoint.
- 12:00 noon
The Noon Curveball is released
Every team receives its track-specific constraint at exactly 12:00 noon IST.
- 12:00–3:30 PM
Reassess, adapt and test
Start a fresh agent session, reconstruct the earlier work and assess the impact before editing.
- 3:30–4:00 PM
Final verification and submission
Documentation, verification and submission.
- 4:00 PM
Code freeze
Submission deadline. No further implementation.
- 4:00–5:00 PM
Demonstrations and judging
Show the product, the decisions behind it, and the evidence that your adaptation holds up.
Meet the judges
Four practitioners score every project — from the platforms the day is built on and the teams building with them. They will ask you to explain your own code.
Debasis Bhattacharya
Director, Field Engineering
Databricks
LinkedIn— opens Debasis Bhattacharya’s profile in a new tab
Meet the mentors
Mentors are on the floor all day. Pull one over when you are stuck, before you have burned an hour on it.
Judging
Judges look at both the working product and the development evidence behind it. A small, well-tested product can score strongly. Every project is scored on the Entire rubric; the Best Use of Databricks rubric is scored separately, and only for teams that opt in.
Entire
100 points- Problem and innovation20
- Technical implementation25
- Response to The Noon Curveball15
- Use of Entire Checkpoints15
- Use of Entire Graph15
- Demonstration and future potential10
Judges will look at both the working product and the development evidence behind it.
Best Use of Databricks
100 points- Meaningful use of Databricks30
- Working implementation and technical reliability25
- User value and product decisions20
- Data quality, provenance and responsible use15
- Response to The Noon Curveball10
Scored only for teams that opt in. It does not change the Entire score or the results of the main challenge. If scores are tied, compare meaningful use of Databricks first, then working implementation and technical reliability.
Explain your own code, show meaningful tests and make the result reproducible. Be ready to explain the implementation and its tradeoffs without asking your coding agent to answer for you.
Prizes
Every award in the Entire main challenge is paid in Amazon vouchers. The awards recognise the top three eligible teams, and amounts are per winning team, not per participant.
₹1,00,000
in Amazon vouchers
1st place
₹75,000
in Amazon vouchers
2nd place
₹25,000
in Amazon vouchers
3rd place
Every award in the Best Use of Databricks category is paid in Amazon vouchers. These awards are additional to Entire’s prize pool: a team may win one Entire placement and one Databricks placement, and each team may receive at most one Databricks placement.
₹25,000
in Amazon vouchers
1st place
₹15,000
in Amazon vouchers
2nd place
₹10,000
in Amazon vouchers
3rd place
Prize stacking is allowed. One project, submitted once, is assessed for both — a team can take an Entire placement and a Best Use of Databricks placement for the same build.
Submission checklist
Every team submits its project name, selected Entire track, team details, repository URL, final commit SHA, demo access, setup and test instructions, and a BUILDATHON.md file.
Your BUILDATHON.md should explain the user and problem, original plan, dependencies and prior components, pre-noon state, assigned Curveball, adaptation, verification, limitations and next steps. Include links to the evidence judges need.
Every team also submits its Entire mirror or project URL, relevant Checkpoint links and Graph findings.
Additional fields for teams entering Best Use of Databricks
- Opt in to Best Use of Databricks: yes or no.
- Databricks products or capabilities used.
- The core function Databricks performs in the product.
- Databricks application, workspace, endpoint or demonstration URL.
- Relevant repository paths and reproduction instructions.
- Data sources, permissions, snapshot dates and limitations.
- How The Noon Curveball affected the Databricks implementation.
Keep credentials out of the repository, documentation, demo recording and submission form. Arrange judge access through the approved access process.
What to bring
Prepare before the day.
- Bring a laptop, charger and the development tools your team can use comfortably.
- Check access to GitHub and your chosen coding agent. Complete a practice Entire setup in a separate disposable repository.
- If entering Best Use of Databricks, check your Databricks account, authentication and ability to deploy or demonstrate the capability you plan to use in a separate practice project.
- Read the selected Entire track, submission requirements and Code of Conduct. Identify your user, risks and smallest useful scope.
- Research, sketches, tool installation and planning are allowed beforehand. Do not implement your competition project before the official start.
- Use only permitted data. Do not put personal information, customer data, private production material or credentials into prompts, repositories or captured development history.
Competition implementation begins at 9:00 AM on 6 September. Previously completed projects and prebuilt feature branches are not eligible.
FAQ
BTW Buildathon 2026 is a one-day, builder-first event focused on turning meaningful developer problems into working products. Every team selects one of Entire’s three original tracks and receives a mandatory Noon Curveball at exactly 12:00 noon.
Take one with your name on it.
Drag it
Every badge is the same until someone’s name is on it. Put yours on, and take it with you.
It renders in your browser and downloads straight to your machine. Nothing is uploaded and nothing is stored — this page has no idea what you typed.
Bengaluru Tech Week
September 1-6, 2026ಬೆಂಗಳೢರು
Bengaluru Tech Week 2026 runs September 1-6, 2026 across Bengaluru. The week brings meetups, workshops, demos, founder gatherings, office hours, and community events into one citywide calendar for founders, engineers, designers, investors, operators, researchers, students, and communities building from Bengaluru for India and the world. Whether you're building your first startup, scaling a global product, exploring AI, investing in the future, or simply curious about what's next, BTW is where the conversations begin.
The week talks. The last day builds.
Sunday, 6 September 20269:00 AM–5:00 PM IST