Winners of the hackathon
CODERHACKx 2023
Calling all innovative student coders and developers to join
Position
Name
1
Srinivasan Sivakumar
2
WeavInsight
3
Rhino Partners
| Position | Name |
|---|---|
| 1 | Srinivasan Sivakumar |
| 2 | WeavInsight |
| 3 | Rhino Partners |
DAY
00
HR
00
MIN
00
SEC
00
#chatbots #ai #coders #developers #sff2023
Organized By
Supported By
About Us
Diving into the Digital Revolution
In a world redefined by the digital revolution, we've evolved from personal computers to the limitless realm of AI. This transformation has touched every facet of our lives, from communication to education and our environment. However, it's not without its unique set of challenges - from cybersecurity threats to digital addiction, misinformation, and the digital divide.
The digital age holds immense promise, but it's our collective responsibility to adapt and govern it wisely.
How can we redefine the future of AI, voice technology and chatbots and bring it to the next level? This hackathon aims to drive new solutions and break boundaries for good.
This official SFF 2023 hackathon is organised by Elevandi and APIX, as part of efforts to advance learning, education and the future of work in the applications AI in Financial Services.
Problem Statements
Introducing CODERHACKx 2023. X marks the space. What will you fill it up with? Race against time and be part of the solution! Join us in solving tech challenges for good, as we aim to elevate the capabilities of Large Language Models (LLMs) within real-time voice-dialogue applications and more.
Are you ready to shape the future? Don’t miss out.
Problem Statement #01
Implement Large Language Models (LLMs)
When deploying Large Language Models (LLMs) in voice-dialogue applications, long-term interaction is the norm, thus requiring continuous input-output capabilities. However, current LLM implementations do not support continuous inputs, only continuous output. The goal is to design and validate a streaming LLM that can achieve continuous input and real-time continuous output.
- Local LLM (of more than 1B parameters) processing using a single GPU.
- Capable of accepting infinite or large input token limits (such as 128K tokens).
- Inputs are in a continuous stream, and the number of bytes added each time is not fixed.
- The implementation should immediately pass the input into the model (considering the actual situation of the voice scenario, a very small buffer, such as no more than 128 bytes, can be retained).
- As soon as a specified terminator (i.e. ) appears in the input stream, the model should start outputting as soon as possible, and the output throughput should not be lower than that of the original LLM model.
- After a terminator appears in the input stream, if there is new incremental input, it should be treated as the next request for processing, and the context before the terminator should be considered when processing the next request.
- The maximum interval between terminators should be similar to that of the current popular LLMs, such as 16KB (about 4K tokens). If the continuous input exceeds the maximum interval but no terminator appears, the previous content can be ignored.
Data for the problem statement is available at AWS S3 bucket
AWS Bucket Name: ps1and2-coderhackx-2023
AWS Access Key ID: AKIAR5ZTUIZKIVFNM654
AWS Secret Access Key: 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
Problem Statement #02
Concurrently Processing Multiple LLM Tasks
Under the requirements and conditions of Problem Statement #01, we need to use a single LLM model on a single GPU to concurrently process multiple LLM tasks. The goal is to simultaneously process multiple LLM tasks under given GPU resources and keep the response time for each task within an acceptable range.
Data for the problem statement is available at AWS S3 bucket
AWS Bucket Name: ps1and2-coderhackx-2023
AWS Access Key ID: AKIAR5ZTUIZKIVFNM654
AWS Secret Access Key: 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
Problem Statement #03
Branch Optimisation
Unable to Forecast Demand in Branch activity and optimize branch load for customers booking appointments.
Unable to recommend solutions to customers to eliminate or deliver through self-serve platforms.
Solution Required:
Maximize operational efficiency by analyzing data to improve staffing, resource allocation, and customer service at their physical branches
Benefits to clients (Banks)
- Optimize staffing
- Increase branch profitability
- Reduce Wait-times
- Cost savings
Expected Output from solution provider
• Recommended approach to solve the problem
• Solution architecture
• Proof of concept
• Auto-generated flat files or API that can be integrated with application
Synthetic data required
- Branch level history data
- Transactions history data
Data for the problem statement is available at AWS S3 bucket
AWS Bucket Name: ps-3-coderhackx-2023
AWS Access Key ID: AKIAR5ZTUIZKMPUCX2E5
AWS Secret Access Key: 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
Prize
Unlock the path to winning exciting prizes in the world of coding innovation!
All Participating Teams
Complimentary SFF Delegate Passes (worth S$1000) and an exclusive CODERHACKx 2023 t-shirt!
Top 3 Finalists
Pitch on the Talent Stage at the Singapore FinTech Festival; Networking meeting with FinTech leaders from the private and public sectors, including Sopnendu Mohanty, Chief FinTech Officer of the MAS and complimentary SFF 2024 tickets!
Winning Team
Networking meal with the folks from the Monetary Authority of Singapore (MAS) AI office and a podcast appearance on Hatch & Hustle FinTech podcast show!
Eligibility
Calling all student coders and developers! Got a build or solution that can tackle any of our challenge statements and a team of 2-5? Awesome! We’d love to have you on board. For student coders, just make sure you’re 18 or older to join in on the fun. Let’s code, collaborate, and make things happen together!
FAQ
Am I eligible?
We welcome all participants interested in building technological solutions to address the identified problem statements and who have a proven track record of delivery – whether that is writing software, designing user experiences, or product management. Participation is expected in the form of teams (typically upto 5 members).
Can I apply to more than one problem statement?
Yes, your team can submit one proposal per problem statement Incomplete entries may be disqualified.
Are there any other requirements?
You will have to agree to the terms and conditions of the Codehack2023 and the APIX Terms of Use at the point of registration.
How to solve the problem statements?
The problem statements should be solved with the data provided for each problem statement.
How will the winning proposals be selected?
Proposals & POCs will be reviewed and three proposals will be shortlisted by our expert panel. The three finalists will have to make a presentation & POC (7 to 10 minutes) in front of the judges on 17th Nov, 2023 at the SFF – Talent Zone between 4 PM – 5 PM. The winner of the contest will be judged by the panelists present on the Demo Day.
How do I register and submit my proposed solution(s)?
Refer to the section "How to Register" below.
Who owns the code developed during the hackathon ?
Teams/ Individuals have full ownsership of the solution proposed and any prototype that they build and host on the APIX Platform. The host of the hackathon and APIX have no ownership of your IP.
How do I get technical support?
To get technical help you can write to support@apixplatform.com