Skip to content
ProjectsWritingAboutCVLinks
Hire me
HomeProjectsWritingAboutCVLinks

Have a useful problem?

Let's make the next
thing worth inspecting.

© 2026 Abdul WahidBuilt around evidence, not hype. Privacy
ABDUL WAHID
Home/Projects/Plan Banao

Case study, Applied AI

Plan Banao
Hackathon V2 · Agent system

An agent system built in one day at MTM Hackathon '26 that reads a pack of scattered messages — emails, WhatsApp texts and notices in English and Roman Urdu — and produces a complete, validated plan for a 4-hour hackathon: venue, schedule, food, judging and budget. The language model only reads; every decision and figure is computed and checked in code.

Built at MTM Hackathon '26 (V2), GDG On Campus CUI Wah — 18 Nov 2026.Applied AI
All projects
Visit

Context

The problem

Event details arrive scattered across messages that contradict and update each other. A useful plan has to reconcile them, respect every hard constraint, and stay valid when a plausible-sounding message is wrong.

My role: Team lead and presenter, Team AXIONS (with Abdul Moiz and Jawad Hassan), GDG On Campus CUI Attock.

Implementation

How it works

The model reads

One extraction call turns raw message text into structured facts, each with its source message and date. A later message from the same sender supersedes the earlier one, and item names are matched flexibly so a correction replaces an entry instead of double-counting it.

The code decides

Teams, judges needed, people to feed, the cheapest venue that fits, the food tier, a schedule that sums exactly to the event length, and the full budget are computed in plain code, so every figure is exact and auditable.

Checks guard the plan

Six validation checks run after every build. An update that breaks a previously passing check — or imposes a quantity below a hard requirement — is rejected, the plan stays valid, and the reason is printed in the trace.

Implementation highlights

  • One command runs the whole system: the model extracts structured facts, and all planning arithmetic is deterministic code
  • Six-check validator — budget, venue fit, food, judges, schedule, coverage; a message that would break a passing check is rejected with the reason printed
  • Transactional updates: a new message is merged into a copy of the fact base first and rolled back if the plan would become invalid
  • The same engine ported to a full web app: plan dashboard, what-if simulator, agent chat and clickable decision evidence

Evidence

What can be verified

The points below summarize the public repository and owner-provided CV where noted. Read the source for implementation details and the evaluation setup before drawing further conclusions.

Sample pack resultRs 547,400 of Rs 650,000

Plan produced from the brief's sample pack (177 people, Hall A); all six validation checks pass.

Runtime≈5 seconds

Full run against the brief's 2-minute limit; updates use at most 2 model calls against a limit of 5.

Tests23/23 offline checks

Regression suite runs against a stub model: sample pack, budget change, the same change twice, and the plausible-but-wrong message (rejected, plan unchanged).

Modelgpt-oss-20b via Groq

One small model for every call, as the brief's small-model bonus required.

Boundaries

Limitations and next questions

  • On the unseen judging pack the venue check failed honestly (5 of 6): the step-free access requirement did not textually match the venue's ramp-access wording, so a cheaper non-compliant hall was selected and the validator flagged it.
  • The system plans one event shape — a single-day hackathon. It is a one-day hackathon build, not a general event-planning platform.
Evidence-first note: a live demo demonstrates an interface and workflow; it does not by itself validate model quality, security, or real-world impact.

Technical overview

PythonGroqJavaScriptVercel

Project links

Inspect the original implementation and, where available, the live application.

Repository

Live demo

Last content review: 14 October 2026. Details may change as project repositories evolve.

Continue exploring

Related work.

Machine Learning

WhyRisk

A live explainable healthcare risk platform: a LightGBM diabetes risk model trained on 70,692 health records (BRFSS), with patient‑level SHAP explanations and role‑based access for patients, doctors, and admins.

Machine Learning

Customer Churn Prediction

A supervised-learning study comparing classification approaches for identifying customers who may leave a service.