Banandre - Page 20

ARTIFICIAL INTELLIGENCE (406)DATA ENGINEERING (110)ENGINEERING MANAGEMENT (56)ENTERPRISE ARCHITECTURE (35)PRODUCT MANAGEMENT (27)SOFTWARE ARCHITECTURE (190)SOFTWARE DEVELOPMENT (213)TECH (1)

Page 20 of 70

AI engineering

Featured

AI Engineering Is Just Data Engineering in Disguise

AI engineering roles are largely rebranded data engineering jobs involving probabilistic data flows, prompt engineering as inputs, and embeddings as features. This perspective demystifies AI development and empowers data engineers to transition without panic or retraining.

#AI engineering#career transition#Data Engineering...

b-tree

Bypassing SQL for 60% Performance Gains: Brilliant Engineering or Dangerous Heresy?

Inside SNKV’s controversial approach to SQLite optimization, direct B-Tree API access that eliminates SQL overhead while preserving ACID guarantees

#b-tree#key-value-store#sqlite...

data-curation

The 1.8M-Parameter Language Model That Questions Everything We Know About Scale

An enthusiast’s journey training a minimal-scale LLM from scratch reveals how architectural innovation and obsessive data curation can squeeze GPT-2 level quality into 25MB.

#data-curation#from-scratch#retention-mechanism...

coding models

Qwen3 Coder Next: The Sub-60GB Model That Makes Cloud APIs Look Overpriced

Qwen3 Coder Next delivers 70.6% SWE-Bench performance with only 3B active parameters, running comfortably under 60GB and finally making local AI coding assistants genuinely usable for interactive development.

#coding models#llama.cpp#local AI...

gguf

The 3D Visualizer That Exposes How Little We Understand About Our Local AI Models

A developer’s rough GGUF visualizer reveals a critical gap: we’re running powerful quantized models with virtually no tools to inspect their internal mechanics, forcing a confrontation between AI democratization and model opacity.

#gguf#mechanistic-interpretability#model-interpretability...

notebooks

Beyond #NotebookEverything: The Production Engineering Debt You Didn’t Know You Were Building

Jupyter notebooks revolutionized data exploration, but shipping them to production creates a silent reliability tax that compounds faster than your data grows. Here’s why the tools that accelerate development actively undermine operational excellence.

#notebooks#production-pipelines#reliability

data-governance

The WHERE Clause Documentation Gap: Where dbt’s Transparency Promise Breaks Down

#data-governance#dbt#documentation

duckdb

DuckDB in Production: The Embedded Database Challenging Enterprise Data Dogma

Real engineering teams are running DuckDB in production, cutting costs by 70% and outperforming Spark clusters. But is an embedded analytics database really ready for enterprise workloads, or are we just tired of overcomplicated data stacks?

#duckdb#motherduck#production

AI Agents

Your Next System Architecture Will Be Designed by 16 AI Agents. Here’s What We Learned Spending $20,000 to Build a C Compiler.

Anthropic’s latest experiment reveals how AI agent teams are rewriting the rules of system design, delivering a 100,000-line C compiler in Rust while exposing the brutal realities of autonomous software development.

#AI Agents#Compiler Engineering#Opus 4.6...

backend-services

The Self-Improving Code Apocalypse: Why GPT-5.3-Codex Makes Backend Services Both Obsolete and Essential

GPT-5.3-Codex debugged its own training and now promises to replace backend development. But the real story isn’t about replacement, it’s about how AI-generated systems create architectural chaos that only stricter platforms can solve.

#backend-services#generative-programming#gpt-5.3-codex

database

It’s 2026 and Postgres Already Won: The Database War No One Saw Coming

PostgreSQL didn’t just survive the specialized database boom, it assimilated it. Here’s how one database ate the entire data stack and why your polyglot persistence strategy is probably technical debt.

#database#postgresql#Scalability

database

The On-Prem AI Database Dilemma: Oracle’s 26ai and Enterprise Skepticism

Oracle’s 26ai brings AI capabilities on-premises, but enterprises are questioning the resource costs, architectural lock-in, and whether on-prem AI workloads make sense when cloud-native alternatives dominate.

#database#on-prem#oracle