Problem Statement 1 - AI-Powered SQL Assistant
-by
P roblem Statement:
Databases require expertise in schema design, SQL dialects, and query
optimization. Many developers, analysts, and data engineers struggle with:
● Designing a well-structured database schemabasedon business needs.
● Generating DDL (Data Definition Language) statementsto create tables.
● Generating DML (Data Manipulation Language) statementsto insert,
update, or delete records.
● Translating natural language queries into SQLwhilesupporting different
SQL dialects such asTrino and Spark SQL.
● Incorporating user feedbackon query correctness andrefining future
query suggestions.
Y our Challenge:
Build anAI-powered SQL assistantthat can:
● Suggest an optimal database schemabased on naturallanguage
descriptions for OLAP use cases.
● Generate correct DDL and DML statementsfor the requestedschema.
● ProvideSQL completion assistance.
● Translate English text into SQL queriesinTrino orSpark SQL dialects,
based on user selection.
● Execute queriesusingTrino or Spark SQL as executionengines.
● Incorporate user feedbackto improve query generationand schema
recommendations over time.
Technical Requirements:
● Dynamically generate SQL statements based on the chosen dialect
(Trino/Spark SQL).
● Use a JDBC-compatible database or a mock execution environment for
validation. Implement feedback learning to refine schema and query
suggestions over time.
● Provide aninteractive UIwhere users can:
○ Input schema requirements
○ Generate and edit SQL statements
○ Execute queries (optional)
● Provide feedback
Utilizefree and preferably open-source AI modelsas required.
Bonus Features (For Extra Points!)
Suggest astorage-optimized and performance-efficientschema.
Leveragespecialized functions based on the selectedSQL dialect.
Evaluation Criteria
1 . Accuracy
Generated SQL statements should be relevant and correct.
2. Functionality
● Proper use ofdialect-specific functions(e.g., UNNESTfor Trino vs.
EXPLODE for Spark SQL).
● Correct usage ofPARTITION BY / CLUSTER BYetc intable
definitions.
3. SQL Translation Accuracy
● Proper use of SQL constructs, includingaggregations,joins, filters,
limits, and ordering
● Queries should beboth syntactically and semanticallycorrect.
4. Performance
● Generated DDL statements should beoptimized for performance,
incorporating best practices like partitioning
● Translated SQL queries should beoptimized for efficientexecution
5. User Interface
The UI should beresponsive, simple, and user-friendly.
P roblem Statement 2 - Metastore Viewer for Parquet,
Iceberg, Delta & Hudi Tables on S3
-by
B ackground
E6data leverageslakehouse formatson object storesand requires an efficient
way to explore metadata without relying on traditional metastore services. While
many metastores allow registeringParquet, Iceberg,Delta, and Huditables, most
do not support metadata exploration when only astoragelocation(such as an S3
bucket containing Iceberg tables) is provided.
Problem Statement
Develop aweb-based metastore viewerthat enablesusers to provide anS3 (or
Azure Object Store / MinIO) bucket pathcontainingParquet, Iceberg, Delta, or
Huditables and retrieve structured metadata, including(but not limited to):
● Table Schema: Column names, data types, partitionkeys
● Table Properties: Format-specific configurations (e.g.,Iceberg’s manifest
files, Delta’s transaction logs)
● Partition Details: Partition columns, pruning statistics
● Versioning & Snapshots: Iceberg’s [Link] ,Delta’s checkpoint files
● Key Metrics: File sizes, row counts, storage optimizations
Requirements
1. Backend
● Readlakehouse file metadatadirectly from S3 (orother object stores)
● Parse metadata files (e.g., Iceberg manifests, Delta logs) to extract table
details
● Expose a RESTful API for the frontend
2. Frontend
● Accept S3 (or other object store) paths to fetch metadata
● Provide an interactive UI to browse schemas, partitions, and snapshots
● Display sample data from the table
● Visualize table evolution and statistics
3. Tech Stack
● Any library if used must be either open source or freeware for enterprise
use.
Bonus Features
● Trino Integration: Execute queries directly on discoveredtables
● Metadata Comparison: Track schema changes over time
Evaluation Criteria
● Functionality: Ability to parse and display metadatafrom object stores
● Usability: Intuitive and user-friendly UI
● Performance: Efficient handling of large and evolvingdatasets
● Innovation: Features such as snapshot comparison andchange
summaries are encouraged
Expected Outcome
A metastore viewerthat allows users to exploreIceberg,Delta, and other
lakehouse table metadata stored in object stores,making it easier to manage and
analyze large-scale table formats without requiring a dedicated metastore
service.
Problem Statement 3 - Job Portal Web Application
B ackground
Many job matching platforms are locked behind paywalls, creating financial
barriers for both job seekers and recruiters. As a result, job seekers struggle to
find opportunities, and recruiters face challenges in sourcing the right talent
efficiently. This hackathon aims to address these issues by developing a
free-to-use job matching platform that simplifies hiring and career growth for
everyone.
O bjective
The goal is to create an accessible and efficient job matching platform that
eliminates financial barriers, ensuring free access to job seekers and recruiters
while optimizing the hiring process.
K ey Features
For Job Seekers:
● Profile & Resume Management:Create profiles and uploadresumes easily.
● Smart Job Search:Find jobs using filters like location,experience, and
industry.
● Application Tracking:Monitor job application statusesin real-time.
For Recruiters:
● Free Job Posting:Post and manage job listings atno cost.
● Candidate Search:Filter candidates by skills, experience,and
qualifications.
● Application Dashboard:Review and manage applicationsefficiently.
E xtra Features
For Job Seekers:
● Resume Builder: Create and customize resumes using an integrated tool.
● Interview Scheduling: Manage appointments with a built-in calendar.
● Job Alerts & Notifications: Get real-time updates on job postings and
applications.
For Recruiters:
● Automated Communication: Send real-time updates to candidates via
email.
● Collaborative Hiring Tools: Enable teamwork for managing listings and
candidates.
● Analytics & Performance Metrics: Gain insights from job post engagement
and candidate activity.
B y removing financial constraints, this platform will bridge the gap between job
seekers and recruiters, ensuring a smooth, cost-free hiring experience for all
stakeholders.
J udging Criteria
Novelty / Innovation – The platform should introduce unique features or
creative solutions that differentiate it from existing job matching platforms.
Usability (Ease of use: UI, UX) – The user interface and experience should be
intuitive, ensuring seamless navigation for both job seekers and recruiters.
Implementation (Functionality) – The platform should be fully functional, with
core features effectively implemented and working as intended.
Presentation Skills – The team should clearly articulate their ideas,
demonstrating the platform’s value, features, and impact effectively.
Adherence to Hackathon Timelines – The project should be completed within
the given time frame, meeting all submission deadlines.
P roblem Statement 4 - AI for Smart and Adaptive
Systems
- by
B ackground
Artificial Intelligence (AI) is transforming various industries by enabling smart,
adaptive, and automated systems. From personalized learning experiences to
AI-driven business intelligence, the demand for AI solutions that enhance
efficiency, creativity, and decision-making continues to grow. However, many AI
applications remain siloed, focusing on specific domains rather than offering
holistic, multi-purpose solutions. This hackathon seeks to explore AI’s potential
across various domains, encouraging participants to develop smart and adaptive
systems that address real-world challenges.
O bjective
The goal is to build AI-powered solutions that can intelligently adapt to user
needs, improve workflows, and enhance creative expression. By focusing on
smart and adaptive systems, this track encourages participants to innovate
across multiple domains, including education, business, software development,
and creative content generation. Participants are free to choose one or more
domains from the given focus areas.
Key Focus Areas
● Personalized Learning – AI-driven tutors that adapt to student progress,
automated exam preparation assistants, and accessibility tools like
text-to-audio summarization.
● AI for Developers – AI-powered code completion, automated bug
detection and fixing, and intelligent documentation generation for
improved software development workflows.
● Business Intelligence – AI-driven financial insights, predictive analytics
for decision-making, and smart automation tools for enterprises.
● Creative AI – AI-generated music, automated video editing assistants,
AI-powered storytelling or meme generation.
B y integrating AI across these domains, participants will have the opportunity to
create versatile, impactful solutions that showcase the true potential of smart
and adaptive AI systems. The challenge is to build innovative, functional, and
user-friendly AI applications that address real-world problems while
demonstrating technological excellence.
Evaluation Criteria
● Novelty / Innovation – The AI solution should introduce unique features
or creative approaches that differentiate it from existing technologies.
● Usability (Ease of use: UI, UX) – The user interface and experience should
be intuitive, ensuring seamless interaction and accessibility.
● Scalability – The AI solutions should be designed to efficiently handle
growing datasets, user interactions, and evolving demands without
compromising performance.
● Implementation (Functionality) – The solution should be fully functional,
with core features effectively developed and working as intended.
● Presentation Skills – The team should clearly communicate their ideas,
demonstrating the solution’s value, features, and impact effectively.
● Adherence to Hackathon Timelines – The project should be completed
within the given timeframe, meeting all submission deadlines.