Skip to content
View Spla4sH's full-sized avatar
6 followers · 15 following

Block or report Spla4sH

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Spla4sH/README.md
Sebastian Roesch - SAP Consultant and Developer on S/4HANA

Portfolio



Featured Work

A reinforcement learning agent beats Bowser in level 8-4
An agent that never saw anything but pixels finishes Super Mario Bros.

Super Mario with Deep RL · Live demo
All 32 levels solved, 20 of 20 greedy runs to the flag each. Along the way I found five metrics that measured something other than what was meant - three of them in my own code.

GitOps Lab
ArgoCD App-of-Apps on a kind cluster, running end to end: ingress, cert-manager, sealed-secrets and monitoring, with CI building the images to GHCR.

Energy Load Forecasting
Rolling-window backtesting for building load: LightGBM, NHiTS and TSMixer measured against a robust seasonal baseline.


Tech stack

Currently Learning

Since 2026 I work in-house as an SAP consultant and developer on S/4HANA - the road there went through SAP test automation at Haufe Group.

Topic Where I'm at
ABAP learning it on the job
CDS Views in use in my project
OData building it up
RAP up next

GitHub stats   Most used languages
Contribution streak

Contribution Snake

Contribution snake
Profile views

Pinned Loading

  1. mario-reinforcement mario-reinforcement Public

    Ein RL-Agent lernt Super Mario Bros nur aus Pixeln und spielt es komplett durch: alle 32 Level geloest (Double DQN + PPO), je 20/20 greedy. Go-Explore, Behavior Cloning und Transfer Learning fuer H…

    Python 1

  2. darts-energy-forecasting darts-energy-forecasting Public

    Rolling-window backtesting pipelines for building energy load forecasting using Darts (LightGBM, NHiTS, TSMixer, GlobalNaiveSeasonal)

    Python

  3. autots-forecasting autots-forecasting Public

    AutoTS Load Forecasting

    Python

  4. react-webshop react-webshop Public

    React single-page webshop with product search, cart, i18n and theming — HFU React course project.

    JavaScript

  5. Fraunhofer-Softwareprojekt Fraunhofer-Softwareprojekt Public

    JavaScript