Hey, I'm Manas β I build systems that turn raw, messy data into something you can act on.
Recent work spans behavioral analytics (Sentinel β detects multi-week attack patterns and outputs auditable dry-run response plans) and market analysis (Market-Oracle β uncertainty-aware signal interpretation across timeframes). On the data side, I've shipped Streamlit dashboards like Retail-analytics-pro that ingest raw sales data, auto-normalize schemas, and surface forecasts.
I work mainly in Python for data/ML pipelines and TypeScript when a project needs a real frontend or backend-for-frontend layer β happy to go full-stack when the problem calls for it.
- π± Active open-source contributor β merged PRs across community projects
- π Currently exploring: applied ML for decision systems (forecasting, anomaly detection, signal interpretation)
- π¬ Open to collaborating on data + AI tooling, especially anything analytics-adjacent
- π Data Engineering & Analytics β schema normalization, ETL pipelines, forecasting dashboards (Retail-analytics-pro, ecommerce-sales-eda)
- π‘οΈ Behavioral Security Systems β multi-week attack-pattern detection with auditable, dry-run response plans (Sentinel)
- π Market Signal Interpretation β uncertainty-aware analysis across multiple timeframes (Market-Oracle)
- π€ Applied ML for Decision Systems β forecasting and anomaly detection as a general toolkit, not just one-off notebooks
- π Full-stack tooling β TypeScript backend-for-frontend layers when a project needs a real UI on top of the data work
| Project | What it does |
|---|---|
| π‘οΈ Sentinel | Detects multi-week behavioral attack patterns and outputs auditable, dry-run incident response plans |
| π Market-Oracle | Uncertainty-aware market signal interpretation across multiple timeframes |
| π Retail-analytics-pro | Streamlit dashboard that ingests raw sales data, auto-normalizes schemas, and surfaces forecasts |
TypeScript ββββββββββββββββββββββββββββββββββββββββ 50%
Jupyter NB ββββββββββββββββββββββββββββββββββββββββ 29%
JavaScript ββββββββββββββββββββββββββββββββββββββββ 13%
Python ββββββββββββββββββββββββββββββββββββββββ 8%
