Templates

Beamer Singapore

Preview

Beamer Singapore

Singapore theme --- subtle purple accents with small section dots at the top. Clean and understated.

Category

Presentation

License

Free to use (MIT)

File

beamer-singapore/main.tex

main.texRead-only preview
\documentclass[10pt]{beamer}
\usetheme{Singapore}
\usecolortheme{orchid}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{amsmath,amssymb}
\usepackage{graphicx}
\usepackage{booktabs}
\usepackage{hyperref}

\title[Clean Data Pipelines]{Clean Data Pipelines \\ From Prototype to Production}
\subtitle{A Case Study}
\author{First Last}
\institute[Example Data]{Data Engineering \\ Example Corporation}
\date{\today}

\begin{document}

\frame{\titlepage}

\begin{frame}{Outline}\tableofcontents\end{frame}

\section{Why Pipelines Fail}
\begin{frame}{The Common Failure Modes}
\begin{itemize}
  \item Silent schema drift across upstream sources.
  \item No observability at transform boundaries.
  \item Brittle orchestration built on cron and hope.
  \item Data that is fresh enough for dashboards but wrong for models.
\end{itemize}
\end{frame}

\begin{frame}{A Real Example}
We found that 43\% of our model-training failures traced back to one upstream
table that quietly began receiving nulls in a previously non-null column.
\end{frame}

\section{Design Principles}
\begin{frame}{Principle 1: Contracts}
Every input and output of every transform is a declared, versioned schema.
Changes require a compatibility check in CI.
\end{frame}

\begin{frame}{Principle 2: Observability}
Every job emits row counts, null rates, and key cardinalities to a metrics store.
We alert on anomalies, not on totals.
\end{frame}

\begin{frame}{Principle 3: Idempotence}
Every pipeline can be replayed for any day in the last 90 without double-counting.
\end{frame}

\section{Results}
\begin{frame}{Before and After}
\begin{table}
\centering
\begin{tabular}{lrr}
\toprule
Metric & Before & After \\
\midrule
Pipeline failures / week   & 12 & 2 \\
Time to detect incident    & 6 h & 20 min \\
Model re-training failures & 43\% & 5\% \\
\bottomrule
\end{tabular}
\end{table}
\end{frame}

\section{Conclusion}
\begin{frame}{Takeaways}
\begin{itemize}
  \item Contracts + observability + idempotence = boring pipelines.
  \item Boring is good --- boring scales.
  \item Start with the highest-impact table and expand outward.
\end{itemize}
\end{frame}

\begin{frame}{Thank You}
\centering
\Huge Questions?\\[0.8em]
\normalsize [email protected]
\end{frame}

\end{document}
Bibby Mascot

PDF Preview

Create an account to compile and preview

Beamer Singapore LaTeX Template | Free Download & Preview - Bibby