ForeSight — IT Project Risk Prediction (Hackathon)
A 24-hour build: ML risk signals, dashboards, and a deployable monorepo—Top 10 at AI-Manthan.
Problem & industry context
IT portfolios leak budget and timeline risk when early warning signals sit in spreadsheets. PMOs want unified views: anomaly flags, cost overrun probability, and drill-down by squad. Hackathon constraints force ruthless scope—one vertical slice that demos end-to-end value beats a slide deck.
Insight
Risk platforms win on trustworthy data plumbing first (dbt-transformed Postgres, clear metrics definitions) and ML second. FastAPI + Next.js + Docker gives judges a live path from ingestion to dashboard. Monorepos help small teams move fast when boundaries (API vs UI vs analytics) stay explicit.
What I built
Co-built ForeSight in 24 hours: FastAPI backend, Next.js 15 frontend, PostgreSQL with dbt, Dockerized monorepo, and ML-driven risk dashboards. Reached Top 10 among 70+ teams at AI-Manthan (AtliQ Technologies). Demonstrated full-stack delivery under extreme time pressure.
Technical approach
Stack and tooling for this work: Python, FastAPI, Next.js, PostgreSQL, dbt, Docker. Topics covered: Hackathon, FastAPI, Predictive Analytics, dbt.
Topics
Related case studies
Microservice Mesh for a Fitness & Wellness Platform
Digiflux Technologies · 2026 – Present
Designing a gRPC microservice mesh and a standalone mobile app for a fitness brand built on wearable data.
AI-Powered Marketplace Content & Search Platform
Digiflux Technologies · 2026 – Present
Owning a platform indexing 517K+ products across 29 marketplaces, with semantic search built for how creators actually search.
An Adobe Photoshop Plugin Real Designers Use
Digiflux Technologies · 2024 – 2025
Shipping a UXP plugin inside Photoshop itself — a different kind of UI problem than a web app.