iiinova — where your digital self keeps evolving

Serg 🤖 AI / Machine Learning / Data Science

@serg-dot-ai.bsky.social

Data Scientist in Ag 🌱 Environmentalist 🦥 Boba drinker🧋Bestselling Author 📚 of "Interpretable Machine Learning with Python" & upcoming "𝘿𝙄𝙔 𝘼𝙄" 🇨🇷→🇺🇸

Posts Replies Media

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 25/01/2025

Key to making a successful agentic system (one of them 😅): match the “Shape Of Business & Customer Systems” - Vin Vashishta @ @datadaytexas.bsky.social

Knowledge Graphs & Ontologies
The Expertise Required To Run The Business
• The Shape Of Business & Oustomer Systems
• The Business's Domain Knonledge
• The Relationship Between Metrics & Outcomes
• The Cornection Between Interventions & Outcomes
• Structural "Causal" Models (Business Standards Alt
Much Lower Than Scientific & Mathematical)

010

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 25/01/2025

Insightful talk on validating LLM generated SQL with mathematical proofs - Ryan Wisnesky @ @datadaytexas.bsky.social

Four ways to compose SQL queries
The mathematics of category theory tells us that there are four ways to compose
SQL queries. This lets us build virtual data fabrics/meshes.

030

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 25/01/2025

“Data Contracts are like a laser pointer and data observability is more like a flashlight” - Mark Freeman @ @datadaytexas.bsky.social

Common Question: Contracts vs.
Observability
Pala Contracis
Data Observability
Prevent specific data quality issues
Highlight data quality trends
Included in CI/CD workflow
Compliments CI/CD workflow
Informed by business logic
Reflects how data captures business logic
Targeted visibility
Broad visibility
Alerts before change
Alerts after change
gable

070

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 25/01/2025

@chiphuyen.bsky.social delivered a masterclass on: 1. Why #AI engineering is product-first as opposed to #MachineLearning engineering 2. What anti patterns we should avoid in AI engineering

ML Engineering:
Data → Product → Model

Al Engineering:
Product → Data → Model →
Competitive advantages

040

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 25/01/2025

“our thinking is outdated” because we must be aware of our limitations - @eevamaija.bsky.social Great topic @ @datadaytexas.bsky.social #datadaytexas

OUR THINKING IS OUTDATED.
Evolutionary beggere:
• Survival required left-brain precision.
• Long-term thinking was secondary.
The Industrial Revolution prioritized efficiency over explontion.
Our brain resists ambiguity, fevoring quick and detal ciented solutions.

041

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 18/01/2025

I have to defend Apple here: — Facebook is the Windows of social media sites - a graceless boomerville. — Apple is like TikTok - elegant and youthful.

"Steve Jobs invented the iPhone, and now they're just sitting on it 20 years later... squeezing people instead of innovating"

020

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 18/01/2025

Just watched “Bad Boys: Ride or Die”. It gets only ⭐️⭐️ from me. Let me predict 2025 is the year of AI writing better sequels than Hollywood!

010

Serg 🤖 AI / Machine Learning / Data Science @serg-dot-ai.bsky.social · 07/12/2024

Seem like I’m fashionably late to the party. I’ll be near the snacks, pretending I’ve been here the whole time and avoiding small talk like a pro!

060