Hyper-Local Air Quality Forecasting
Predict street-level air quality 6-24 hours ahead in Tamil Nadu's industrial corridors by fusing TNPCB CAAQMS data with weather, traffic, and industrial activity signals.
More DetailsDeadline
Level 1 Idea Submission Closes In
--
Days
--
Hours
--
Minutes
--
Seconds
Submissions Closed
April 7, 2026 · 11:59 PM IST
About the Hackathon
Applying Data-Centric AI to Real
Tamil Nadu Industry Challenges
The curriculum capstone for the Naan Mudhalvan Advanced AI/ML course. Demonstrate that better data, not just more complex models, can produce measurably better outcomes on real Tamil Nadu industry problems.
Data-Centric AI
Diagnose data quality issues, design enrichment strategies, and show a measurable performance gap between baseline and improved models.
Real Industry Problems
20 challenges from Tamil Nadu's textile mills, power grid, transport, healthcare, agriculture, and industry partners
Teams of 2-5
Open to students enrolled in the Naan Mudhalvan Advanced AI/ML course across Tamil Nadu engineering colleges.
Three Stages
Ideation, prototype development, and a live final event replicating the professional AI/ML project lifecycle.
0
Problem Statements
0
Industry Sectors
0
Competition Stages
0
Max Team Size
Hackathon Timeline
From Idea to Final Stage
Seven milestones across three levels, culminating in a live final event at Velamaal Institute of Technology, Tiruvallur.
March 31, 2026
Level 1 Launch
Idea submission opens. Review the 20 problem statements, select one, and prepare your structured ideation document.
Online
April 7, 2026
Level 1 Deadline
Submit your idea via Google Form by 11:59 PM IST. No coding required -- demonstrate problem understanding and proposed approach.
Online
April 10, 2026
Level 1 Results
Shortlisted teams announced. Top teams advance to prototype development with dataset access.
Online
April 11 -- 22, 2026
Level 2: Prototype Development
Build your data-centric AI prototype. Demonstrate measurable performance improvement over a baseline model using enriched data.
Online
April 27, 2026
Level 2 Results
Finalist teams selected for the final hackathon event based on prototype quality and data-centric approach.
Online
April 28 -- May 8, 2026
Mentorship Sessions
Finalists receive guidance via AI Voice Agents (on-demand, 24/7) and industry expert sessions (for Level 3 shortlisted teams).
Online
May 11, 2026
Level 3: Final Hackathon Event
Live demo and defence before a jury of industry experts, Intel representatives, and TNSDC officials. Winners announced on stage.
Physical
Problem Statements
20 Real-World Challenges
Across 11 Industry Sectors
Each problem is grounded in a documented Tamil Nadu operational context. Choose a challenge, explore the underlying data issues, and develop a data-centric AI solution.
Local LLM Medical Document Intelligence
Build a fully local, privacy-preserving RAG system for Tamil Nadu hospital clinicians under DPDP Act constraints -- zero cloud data exposure, fully on-premises inference.
More DetailsCotton Blend Optimization
Model non-linear cotton blend compositions to predict yarn quality in Coimbatore spinning mills, replacing costly trial-and-error blending with computational optimization.
More DetailsDye Recipe Formulation
Improve first-time-right colour matching in Tirupur textile dyeing by predicting reactive dye recipe outcomes, reducing the 25-45% re-dyeing failure rate.
More DetailsBiomass-Coal Co-firing Optimization
Optimize biomass-coal blend ratios at TANGEDCO thermal plants to predict boiler heat rate and steam generation from variable-quality combined fuel properties.
More DetailsMetal Alloy Composition Prediction
Predict mechanical properties of cast iron and ductile iron castings from melt chemical composition to reduce rejection rates in Coimbatore's 300+ foundries.
More DetailsRubber Compound Property Prediction
Predict physical and performance properties of rubber compounds from formulation recipes to reduce the 2-4 week development cycle for automotive sealing and vibration control.
More DetailsSeasonal Disease Surge Prediction
Predict disease surges (dengue, scrub typhus, gastroenteritis) in Tamil Nadu government hospitals 2-3 weeks ahead using leading indicator signals for proactive resource mobilization.
More DetailsPilgrim Footfall Prediction
Improve footfall prediction at major Tamil Nadu temples (Madurai Meenakshi, Tiruvannamalai, Rameswaram) using proxy signals and Tamil lunar calendar patterns.
More DetailsTamil OTT Content Performance
Correct systematic under-representation of semi-urban, rural, and older audiences in Tamil OTT engagement data so content decisions reflect the full Tamil audience.
More DetailsTNSTC Fleet Electrification Planning
Develop an optimal 15-year fleet electrification strategy for TNSTC's 22,000+ buses across 20,000+ routes, minimizing cost while meeting CO2 reduction targets through 2040.
More DetailsCauvery Water Allocation Optimization
Optimize seasonal Cauvery water allocation across delta districts (Thanjavur, Tiruvarur, Nagapattinam) for three paddy crop seasons under uncertain upstream inflows.
More DetailsDrug Demand Uncertainty Quantification
Model the full probability distribution of outbreak-sensitive medicine demand for TNMSC, especially heavy-tailed scenarios, to set scientifically grounded safety stock levels.
More DetailsWind Farm AEP Uncertainty Estimation
Produce calibrated P50/P75/P90 Annual Energy Production estimates for Tamil Nadu wind farms, accounting for monsoon variability and dense-cluster wake effects.
More DetailsGST Revenue Anomaly Detection
Detect GST anomalies across Tamil Nadu's 15 lakh registered dealers by cross-referencing electricity consumption, freight movement, and employment data against return filings.
More DetailsRAG Knowledge Base Quality Engineering
Audit and improve an 850-article knowledge base to reduce chatbot hallucinations -- treating the root cause as documentation debt rather than an LLM tuning problem.
More DetailsTANGEDCO Transformer Failure Prediction
Predict distribution transformer failures in TANGEDCO's rural network by identifying overloading and degradation patterns in smart-meter load data before catastrophic failure.
More DetailsChennai Metro Causal Impact
Quantify the causal impact of Chennai Metro Rail Phase 1 on residential property values using difference-in-differences methodology, separating the true "metro premium."
More DetailsMTC Bus Frequency Optimization
Develop a reinforcement learning agent that adaptively optimizes MTC Chennai's bus service frequency in real time across 800+ routes using smart card and AVL data.
More DetailsChennai Traffic Signal Timing
Develop a reinforcement learning agent for adaptive traffic signal timing on Chennai arterial corridors, handling mixed traffic with 60-70% two-wheelers.
More DetailsPrizes & Recognition
Each Stage Opens New Doors
Each stage builds on the last from dataset access and hands-on learning, to mentorship and recruiter visibility, to formal certificates and internship opportunities at the finals.
All Participants
Level 1 Completion
- Access to hackathon problem datasets and resources for continued learning
- Hands-on exposure to real-world AI/ML problem statements and structured solution approach . Participation in orientation sessions to support problem understanding
Level 2 Winners
Finalist Recognition
- Certificate of Achievement acknowledging prototype development and Level 3 selection
- Access to mentorship sessions AI Voice Agents and industry expert sessions
- Profile listing on hackathon outcomes page for recruiter visibility
Level 3 Winners
Winner & Runner-Up
- Winner and runner-up certificates jointly issued by TNSDC, Intel, and SL2 with commemorative mementos
- Shortlisted for internship placement with partner organisations and SL2
- Considered for entry to the Intel AI Impact Festival for national-level visibility
Mentorship Model
Two Channels of Expert Guidance
Finalist teams receive structured mentorship from April 28 to May 8, 2026 combining always-available AI guidance with direct industry expert access.
AI-Powered
Conversational AI Voice Agent Mentorship
On-demand access to AI Voice Agents developed by Sustainable Living Lab India available 24/7 without scheduling constraints throughout the mentorship window.
- Technical guidance: data pipeline design, model architecture, evaluation methodology
- Product coaching: solution framing, use-case articulation, real-world applicability
- Pitch preparation: presentation structure, narrative clarity, judge Q&A readiness
Industry Experts
Industry Expert Mentorship
Direct mentorship from domain experts and AI/ML practitioners from problem statement owner organisations, Intel, and SL2's partner network.
- Group sessions: domain context, common pitfalls, and final presentation expectations
- Mentors drawn from problem owners, Intel, and SL2 partner network
Evaluation Criteria
How Submissions Are Judged
Weighted criteria are applied at every stage. The final ranking combines the Level 2 prototype score (40%) with Level 3 live performance (60%).
Problem Understanding
30%
Depth of analysis of the data quality issue and operational context
Proposed Approach
30%
Feasibility and relevance of the proposed AI/ML and data-centric strategy
Innovation
20%
Originality of the data enrichment or methodology approach
Presentation Quality
20%
Clarity, structure, and completeness of the ideation document
Data-Centric Approach
35%
Quality of data analysis, enrichment pipeline, and bias correction methodology
Performance Improvement
30%
Measurable and credible gain of improved model over baseline on test data
Technical Implementation
20%
Code quality, reproducibility, and prototype functionality
Documentation
15%
Clarity of technical report and results communication
Real-World Applicability
30%
How readily the solution could be adopted by the problem-owning organisation
Technical Depth
30%
Sophistication and soundness of the data pipeline and model implementation
Innovation & Impact
20%
Originality of approach and potential for measurable operational impact
Presentation & Q&A
20%
Clarity of demonstration and quality of responses to judge questions
How to Participate
Four Steps to Get Started
1
Read the Problem Statements
Browse all 20 challenges above. Select the one that aligns with your skills and your institution's industry geography.
2
Download the Ideation Template
Use the structured template to document your problem understanding, proposed approach, and data strategy.
3
Form Your Team
Assemble a team of 2-5 students enrolled in the Naan Mudhalvan Advanced AI/ML course at your institution.
4
Submit via Google Form
Fill in your team details, upload your ideation document, and submit before April 7, 2026 at 11:59 PM IST.
Problem Statement Partners
Challenges Sourced From Industry
The problem statements in this hackathon are developed in collaboration with the following organisations, rooting the competition in real industry needs across Tamil Nadu.
