Snigdha Kanchi

Get amplified by AI. Data Leader sharing real AI-native workflows.

[Snigdha Kanchi @snigdata]

@sunrao @bot @SpaceXAI @kinopee_ai @krismatterz Love the swag!

[Snigdha Kanchi @snigdata]

I’m at WeAreDevelopers for the next three days, meeting builders and exploring what’s next in tech. 💻

What should I ask them? Drop your question below and follow along! #WeAreDevelopers

[Snigdha Kanchi @snigdata]

@TAMU @nvidia This is incredible!! Go Aggies!!

[Snigdha Kanchi @snigdata]

Gmail just built an AI agent into your inbox.

AI Inbox finds important emails, creates suggested to-dos, surfaces the related messages and lets you mark tasks complete.

I’ve been testing the beta for a couple of days, and I’m already loving it. ✅

Would you use this?

[Snigdha Kanchi @snigdata]

@covacut @OpenAI This is super exciting! I am an artist and a tech creator! Let’s connect!

[Snigdha Kanchi @snigdata]

@tankots This is incredible!! Congratulations!! 🎉

[Snigdha Kanchi @snigdata]

Meta’s $17.1B settlement is getting attention for the number.

I think the bigger story is the product design precedent.

Meta is introducing default protections for teens, while part of the settlement is conditional on YouTube + TikTok adopting protections of their own.

After 8 years at Meta, my biggest takeaway:

Defaults matter.

A safety feature buried in Settings is very different from a protection built into the product by default.

Could this become the blueprint for how apps are designed for kids?

#Meta #BigTech #TechNews

[Snigdha Kanchi @snigdata]

You do not need an expensive AI bootcamp to start learning AI.
Here are 7 free courses from companies actually building the technology 🧵

1. GOOGLE
Introduction to Generative AI
A quick beginner-friendly introduction to generative AI and how it differs from traditional machine learning.
https://t.co/JDBzT6Xqo3

2. ANTHROPIC
Building with the Claude API
Learn prompting, structured outputs, tool use, and how to build Claude-powered applications.
https://t.co/YRa1L5WgJy

3. OPENAI
AI Foundations
Learn how LLMs work, how to give better instructions, and how to review AI-generated outputs responsibly.
https://t.co/JPwWZcAiFA

4. MICROSOFT
Generative AI for Beginners
An 18-lesson course covering LLMs, prompt engineering, RAG, agents, and AI application development.
https://t.co/i2MCWpMbut

5. AWS
Cloud Quest: Generative AI Practitioner
Learn prompting, model selection, Amazon Bedrock, and practical GenAI concepts through a game-based experience.
https://t.co/6EaeR5xjYz

6. NVIDIA
Generative AI Explained
Learn about generative models, their applications, their limitations, and the technology powering modern AI.
https://t.co/BwIRqkJ7Ub

7. IBM
AI Learning With SkillsBuild
Explore free paths covering AI fundamentals, generative AI, responsible use, and workplace applications.
https://t.co/G9I5th3aTW

Do not collect courses.
Pick one. Complete it. Build one project. Share what you learned.
That is how a free course becomes visible proof of your skills.

Follow @snigdata for practical AI, data, and tech career insights.

[Snigdha Kanchi @snigdata]

Elastic jumped nearly 20% after earnings.

But the interesting part isn’t just the stock.

Nearly $500M quarterly revenue.
15% YoY growth.
1,800+ $100K customers.

And a growing bet that search becomes core infrastructure for enterprise AI.

I spoke with Elastic Principal PM Kapil Jadhav for the latest episode of The Pitch Room.

As AI gets better at generating answers, the ability to find the right context may become even more valuable.

Elastic is one to watch. 🔎

#Elastic #AI #ThePitchRoom

[Snigdha Kanchi @snigdata]

AI agents are getting more powerful. But are we thinking enough about what happens when they get things wrong?

We’re moving from AI that simply answers questions to AI that can take actions, access files, browse websites, execute code, connect to tools, and interact with real systems.

And that changes the risk.

A developer recently reported that Claude Opus 5 mistakenly deleted their Windows user profile while attempting to create a backup. Separately, Microsoft security researchers have demonstrated vulnerabilities in AI-agent frameworks where prompt injection could ultimately lead to code execution.

The takeaway isn’t that we should stop using AI agents.

It’s that more autonomy needs stronger guardrails.

Give agents the minimum access they need. Require human approval before destructive or high-impact actions. Keep sensitive systems isolated. Monitor what agents are doing. And test workflows in low-risk environments before giving them broader permissions.

We’ve spent the last few years asking:

“How capable can AI become?”

As agents become mainstream, we also need to ask:

“How much control should we give them?”

The future of AI isn’t just about building more capable agents. It’s about building agents we can trust with real-world actions.

What’s one action you would never let an AI agent take without your approval?

#AI #AIAgents #AgenticAI #AISafety #ArtificialIntelligence #Cybersecurity #TechTrends

[Snigdha Kanchi @snigdata]

LLMs, skills, and agents explained using a chef:

🧠 LLM: Everything the chef knows
🔪 Skill: A specific thing the chef can do
🍽️ Agent: The chef using both to prepare the meal

Knowledge. Skill. Action.

AI becomes easier when we connect it to the real world.

[Snigdha Kanchi @snigdata]

What does the future of robotics and advanced manufacturing look like? 🤖

At the Plug and Play Microelectronics & Robotics Expo in Austin, one takeaway stood out:

As machines get smarter, data quality becomes infrastructure.

Accurate sensor data. Reliable pipelines. Shared definitions. Real-time monitoring. Human context.

Because whether you’re training a robot, predicting equipment failures, or automating a factory, the intelligence is only as reliable as the data feeding it.

Smarter machines start with better data.

#AI #Robotics #Data #AdvancedManufacturing

[Snigdha Kanchi @snigdata]

The data professional’s toolkit has changed:

YouTube tutorials → AI tutors
Stack Overflow → ChatGPT + Claude
Manual SQL → AI-assisted SQL
Static dashboards → Conversational analytics
Reactive checks → Automated monitoring
Handwritten notes → AI summaries

The tools make us faster. Context and judgment make the work valuable.

[Snigdha Kanchi @snigdata]

5 roles I think will grow because of AI:

Forward Deployed Engineer
AI Product Manager
AI-native Data Engineer
Solutions Architect
AI Governance

The career moat: technical depth + customer understanding + storytelling + business judgment.

Which one would you bet on?

[Snigdha Kanchi @snigdata]

@TAMU Nostalgic memories!!

[Snigdha Kanchi @snigdata]

Your inbox is becoming an AI interface. 📩

AI assistants can help you:

• Prioritize unread emails
• Find messages that need a response
• Extract action items
• Summarize threads
• Prep for meetings

Multiple accounts? Use one primary inbox and selectively forward what you want included.

The shift:

Ask AI → Give AI context → Connect AI to where your work already lives.

Would you connect AI to your inbox?

[Snigdha Kanchi @snigdata]

What infrastructure keeps AI agents running at scale?

In this episode of The Pitch Room, I’m featuring Novita AI. The company offers 200+ AI models, on-demand GPUs, and secure agent sandboxes while processing over one trillion tokens per day.

Hear directly from co-founder and COO Junyu Huang.

#NovitaAI #AIAgents #AIInfrastructure

[Snigdha Kanchi @snigdata]

@huntertcarver @cursor_ai @AntlerGlobal @jake_oshea @ftnabeelah @benln @sunrao This was an amazing event! Glad to have witnessed it first hand!

[Snigdha Kanchi @snigdata]

@TAMU This is amazing!

[Snigdha Kanchi @snigdata]

8 years of building data at Meta taught me something important: AI is only as powerful as the systems, context, and people behind it.

Here are 5 lessons I’m carrying into the AI era:

Good AI starts with good data.
Context matters. Build a “second brain” that preserves the history, goals, and decisions behind the data.
Metrics need shared definitions. If teams define success differently, AI won’t magically create alignment.
Automation without judgment creates problems. AI can scale decisions, but human judgment still matters.
Amplify your impact by sharing knowledge. What you learn can become someone else’s shortcut.

The biggest lesson?

Good data powers AI. Great teams drive impact.

Save this for your next AI or data project, and follow @snigdata for practical lessons on data, AI, and the future of work.

#ArtificialIntelligence #AI #Data #DataEngineering #Analytics #FutureOfWork #TechLeadership #DataLeadership #AITransformation #Meta #SnigData