ARTICLES
Why It’s High Time to Switch from Terraform to OpenTofu | 3 min | DevOps | Nikhil Donthula | KPMG UK Engineering Blog
HashiCorp’s license shift and IBM’s acquisition make Terraform’s future uncertain. OpenTofu, backed by the Linux Foundation, offers a safer, fully open alternative.
Why Scaling a Database Is Harder Than Scaling a Server | 5 min | Data Base | Himanshu Singour | Personal Blog
Servers scale easily with load balancers, but databases face state, consistency, and sharding challenges. This article breaks down why database scaling is fundamentally harder.
TUTORIALS
From Zero to GPU: Building & Scaling Production-Ready CUDA Kernels| 12 min | ML | David Holtz & Daniël de Kok | Hugging Face Blog
A step-by-step guide to writing custom CUDA kernels, integrating them into PyTorch, and sharing them on Hugging Face for production use.
Starting Power BI Deployment Pipelines from Azure DevOps | 8 min | DevOps | Adrian Chodkowski | Seequality Blog
How to connect Azure DevOps with Power BI deployment pipelines using service principals, extensions, and YAML-based CI/CD.
TOOLS
OLake | 7 min | Data Engineering | Olake.io
OLake replicates Postgres, MySQL, MongoDB, and Oracle to Apache Iceberg at up to 64K RPS, with CDC, schema discovery, and a lightweight Docker UI.
Grok Code Fast 1 | 6 min | AI | xAI Blog
xAI introduces a fast, low-cost coding model optimized for IDEs like Cursor and Copilot, free for now via launch partners.
Snowflake Universal Lineage | 8 min | Data Engineering | Snowflake Docs
Snowflake extends lineage beyond its walls using the OpenLineage standard, letting dbt, Airflow, and others send lineage events into Snowsight.
DATA TUBE
How 11x Rebuilt Their Alice Agent: From ReAct to Multi-Agent with LangGraph| 20 min | AI | Sherwood Callaway, Keith Fearon | LangChain
Inside the redesign of Alice, an AI SDR, moving from single-agent ReAct patterns to production-ready multi-agent architectures.
PODCAST
LLM Deployment on Kubernetes with LLMD | 52 min | LLM | Serge Gershkovich | Data Engineering Podcast
Serge Gershkovich shares how teams can model data collaboratively and deploy large language models on Kubernetes.
EVENTS, CONFS, AND MEETUPS
ML in PL Conference 2025 | 1st October | Warsaw
Registration is open for ML in PL 2025, bringing researchers and practitioners together for Europe’s leading ML conference.
PINNACLE PICKS
Your last week top picks:
Data & AI Monitor Report 2025–2026 | AI & Data Strategy | Xebia
Industry benchmarks on GenAI adoption, MLOps maturity, and platform modernization for the year ahead.
Modernizing Sports Betting with Real-Time Data Streaming | 7 min | Streaming Architecture | Mitchell Gray | Ververica Blog
How sports betting platforms use Flink streams to power live odds, fraud detection, and personalization with exactly-once guarantees.
7 Drop-In Replacements to Instantly Speed Up Your Python Data Science Workflows | 6 min | Data Engineering | Jamil Semaan | NVIDIA Developer Blog
Swap Pandas, NumPy, and scikit-learn for GPU-optimized drop-ins like cuDF, CuPy, and cuML for instant speedups.
________________________
Have any interesting content to share in the DATA Pill newsletter?
