> ## Documentation Index
> Fetch the complete documentation index at: https://freelance-4888ed08.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Roadmap

> Future development plans and upcoming features.

This roadmap outlines the strategic direction for the Demand Forecasting system.

```mermaid theme={null}
gantt
    title 2026 Development Roadmap
    dateFormat YYYY-MM-DD
    axisFormat %b %Y

    section Scalability
    Kubernetes Operator       :2026-01-01, 90d
    GPU Support for CatBoost  :2026-02-01, 60d
    Model Registry UI         :2026-02-15, 45d

    section Advanced Modeling
    Deep Learning Candidate   :2026-04-01, 90d
    External Regressors       :2026-04-15, 75d
    Real-time Serving (gRPC)  :2026-05-01, 60d

    section Infrastructure
    Auto-Scaling Inference    :2026-07-01, 90d
    Shadow Mode Deployment    :2026-07-15, 75d
```

## Q1 2026: Scalability & Operations

> [card] Kubernetes Operator
>
> Migrate from Docker Compose to a Helm-managed Kubernetes deployment (EKS).

> [card] GPU Support
>
> Enable `task_type="GPU"` for CatBoost to reduce optimization time by 50%.

> [card] Model Registry UI
>
> Deploy MLflow UI behind OAuth proxy for broader team access.

## Q2 2026: Advanced Modeling

> [card] Deep Learning
>
> Introduce LSTM or Transformer-based model into the A/B testing rotation.

> [card] External Regressors
>
> Integrate holiday calendars and major local event data into the feature store.

> [card] Real-time Serving
>
> Move from Batch Inference to low-latency Online Inference (gRPC).

## Q3 2026: Infrastructure & Resilience

> [card] Auto-Scaling Inference
>
> Horizontal Pod Autoscaling (HPA) for the Model Server based on request volume.

> [card] Shadow Mode
>
> Deploy "Shadow" models that receive traffic but don't return predictions (for safe testing).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.