I create content about deep learning | The DevRel and Deep Learning Dude

I don’t wanna hop on call to learn about or see a demo of what you’re building. Too busy hacking on VLMs, curating multimodal datasets, and contributing to the open source community. 

Harpreet Sahota 🥑

It was great to be back in Boston last week and host another awesome meetup!

Thank you for everyone who showed up, our awesome lineup of speakers for taking time out of their schedules for delivering such informative talks: Anku Rani Hanhui Wang Sage Simhon Sharut G. 

And for Voxel51 for sponsoring!

If you missed the event, checkout the comments for where you can find the recordings of the talks

🤖 Hacker-in-Residence @ Voxel51| 👨🏽‍💻 AI/ML Engineer | 👷🏽‍♀️ Technical Developer Advocate | Learn. Do. Write. Teach. Repeat.

No alternative text description for this image

i built a ComfyUI-inspired visual, node-based workflow editor for FiftyOne

now you can build dataset curation, analysis, and model evaluation pipelines by connecting nodes on a canvas, without writing code

give it a try and let me know what you think, completely open source and all you have to is `pip install fiftyone` and follow the instructions on the readme

everyone is hype about gtc, but the 3d vision conference is this week and i'm much hype for that

i made a repo and a dataset to explore the papers

checkout the repo here:
 https://lnkd.in/gRFz6Kb7

here's a dataset that you can use to explore the papers:
 https://lnkd.in/g2RA87fD

it's been a long, long while since i've done a live stream on linkedin, but it's happening in t-mius 8 minutes!

if you're interested in learning about how to work with video datasets using open source tools, open data, and open models...this is the session for you!

Join us on Feb 26 for Exploring Video Datasets with FiftyOne and Vision-Language Models. Register for the Zoom: https://lnkd.in/gCk7gGPi

Video is the hardest modality to work with — more data, more complexity, and annotation workflows that don’t scale. This hands-on session by Harpreet Sahota 🥑 shows how to understand what’s inside massive video datasets without manually watching thousands of clips.

You’ll learn how to:
🎬 Navigate and explore video datasets in the FiftyOne App and understand dataset structure
🔎 Compute embeddings with Qwen3-VL to enable semantic search, zero-shot classification, and clustering
📝 Generate video descriptions and localize events using vision-language models like Qwen3-VL and Molmo2
🧭 Visualize patterns and trends through embedding projections in FiftyOne
📏 Evaluate model outputs against Action100M’s hierarchical annotations

Perfect for ML engineers, data scientists, and computer vision teams working with large-scale video data who want practical workflows for exploration, curation, and debugging.

#computervision #machinelearning #ai #datascience #videoanalytics #visionlanguagemodels #mlops #opensource

Derek Van de Streek, Swati Bannore Jain and 11 other attendees

Feb 26 - Exploring Video Datasets with FiftyOne and Vision-Language Models

excited to be back in stuttart! if you're in the area and want to speak at the meetup on april 21st, send me a dm

Chief Community Officer @ Voxel51 | Startup Advisor

Are you based in the Stuttgart area? Consider speaking at the upcoming April 21 - Stuttgart AI, ML and Computer Vision Meetup: https://voxel51.com/speak

#computervision #ai #artificialintelligence #machinevision
#machinelearning #datascience #opensource #stuttgart

i've learned a lot from Eric Ma over the years, from the first time I saw one of his talks on principled data science workflows to talks he's given for my communities over the years, and his latest work on getting the most from coding agents.

check out his next talk, and support the SciPy community as well!

Together with my teammates, we solve biological problems with network science, deep learning and Bayesian methods.

As a data scientist, how much of your week is spent on work a coding agent could do?

I once planned multiple days of work for a machine learning experiment last year. This was an experiment that I had put off for an entire quarter because, as a team lead, I didn't have the time to block out the 3-4 afternoons that I thought I'd need. 

Instead, I described the problem to Claude Opus 4.5 via OpenCode, and it wrote the code, ran the training, found issues, switched strategies, and finished the whole thing in 20 minutes, complete with a written report drafted *with* figures referenced.

The data scientists who figure this out now will have an edge.

I'm teaching a workshop on how to do this flavour of agentic data science to raise funds for SciPy 2026's Financial Aid Program. You'll learn:

- Prescriptive prompting for experiments
- Project structure for agents
- Logging and verification strategies to enable automated control
- Report writing with AI
- Staying in control during EDA
- The minimalist version first

These are the same techniques I use in my personal projects and my teammates to compress data science timelines from days to minutes. I think you'll have a chance to learn something good while doing something good!

Details:

- Date: March 27 & April 3, 2026
- Time: 2:00-4:00 PM EST
- Venue: Google Meet (calendar invite will be provided)
- 40 spots available
- $1,000 minimum donation, paid directly to NumFOCUS (tax-deductible), student pricing: $200!

Video recordings available if you can't attend live.

Sign up here: https://lnkd.in/eFjc36Tb. I'm looking forward to seeing what you'll build with these techniques! 

P.S. If you can't make it but share it with others who have the means and would benefit from the workshop, I would deeply appreciate it!

Open article: Agentic Data Science Workshop | SciPy 2026 Fundraiser by agent-assisted-data-science.vercel.app

Agentic Data Science Workshop | SciPy 2026 Fundraiser

We just released a new AWS workshop that I’m really excited to finally share 👇
https://lnkd.in/eiNprTd7

This came from something I keep seeing across autonomy, robotics, and computer vision teams lately.

Most teams are not blocked by models anymore.
They are blocked by data and iteration speed.

So we wanted to show what a practical workflow actually looks like:
🚗 generate synthetic driving data
 🧠 auto label using foundation models
 🔧 fine tune a small vision language model
 ☁️ deploy and iterate using AWS SageMaker
 👀 inspect and improve everything inside FiftyOne

The goal was simple.

Build something real that teams could follow end to end.

Huge credit to Ryan Patrick Sweeney, Mark Vinciguerra, and Derick Thompson for driving this forward, along with great collaboration between Amazon Web Services (AWS), NVIDIA Omniverse, and Voxel51 🙌

Also excited that Voxel51 is now live on AWS Marketplace, which makes getting started a lot easier for teams working on computer vision and robotics workflows!

Personally, I think this is how a lot of Physical AI systems will get built going forward.

Faster loops between data, models, and deployment instead of long waiting cycles.

If you’re working on robotics, AV, or multimodal systems, I’d genuinely love to hear what you’re building.

More coming soon 🚀

#PhysicalAI #ComputerVision #Robotics #MultimodalAI #SyntheticData #AWS #MachineLearning #AutonomousSystems

Open article: GitHub - aws-samples/sample-finetuning-small-vlms-synthetic-data by github.com

Contribute to aws-samples/sample-finetuning-small-vlms-synthetic-data development by creating an account on GitHub.

GitHub - aws-samples/sample-finetuning-small-vlms-synthetic-data

I’m so happy to have my boi Wiz present at one of our meets!

AI @ NVIDIA - 🏗️ Build | 🚢 Ship | 🚀 Share

View Chris "The Wiz 🪄" Alexiuk’s profile

skills.md - you've seen it, you've probably used it (I assume if you're following this LinkedIn account, you're someone who uses tools like OpenCode/etc.

I'm excited to be yappin' about skills at an upcoming Voxel51 / Hugging Face virtual event with a line up that boggles the mind: 

Ben Burtenshaw, Daniel Ávila Arias, and AdonAI Vera.

My presentation will focus on something near and dear to my heart: Skills as documentation. 

The days of documentation aren't over - humans, and agents, still need to understand how things work - but they're changing quickly. Not everyone wants to read a thorough getting started guide that may be built for a specific system (and if it's not, it'll be ~2K words long) Sometimes, you just want to clone a repository - and get cooking. 

Skills help bridge that gap - context for LLMs, but tailored for a human experience. 

Can't wait for the event, come on out and ask any skill questions you have!

looking forward to hosting this workshop tomorrow, join me!

Join Harpreet Sahota 🥑 on Feb 26 for a virtual workshop to learn how to use Facebook's Action100M dataset and FiftyOne to build an end-to-end workflow - https://lnkd.in/gFvwbQni

What you'll learn:

* Navigate and explore video data in the FiftyOne App, filter samples, and understand dataset structure

* Compute embeddings with Qwen3-VL to enable semantic search, zero-shot classification, and clustering

* Generate descriptions and localize events using vision-language models like Qwen3-VL and Molmo2

* Visualize patterns in your data through embedding projections and the FiftyOne App

* Evaluate model outputs against Action100M's hierarchical annotations to validate what the models actually capture

By the end of the session, you'll have a reusable toolkit for understanding any video dataset at scale, whether you're curating training data, debugging model performance, or exploring a new domain.

***********

Want to build better computer vision models? FiftyOne is an open source toolkit from Voxel51 (our Meetup sponsor) that helps you curate datasets, evaluate model performance, visualize embeddings, catch annotation errors, and eliminate duplicate images—all in one place.

“pip install fiftyone” is all it takes to get started - https://docs.voxel51.com/

#computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #opensource

Data Analytics - Turning Coffee into Insights, One Caffeine-Fueled Query at a Time! | Healthcare Data | Financial Expert | Driving Business Growth | Data Science Consultant | Data Strategy

As a data analytics professional, I have a passion for turning raw data into meaningful insights that drive business decisions. With my experience in data visualization, statistical analysis, and data modeling, I can help your organization harness the power of data to optimize operations, improve customer experiences, and achieve your strategic goals.

🔍 Discovering Insights

My expertise lies in exploring complex data sets, uncovering hidden trends and patterns, and providing actionable insights that support informed decision-making. I can help you answer questions like "What are our customer's buying habits?" or "What factors are impacting our sales performance?", and use data to guide your business strategy.

🎨 Data Storytelling

I believe that data storytelling is an art form, and I love using creative and engaging data visualizations to communicate complex information in a way that's easy to understand. Whether it's through interactive dashboards or custom reports, I can help you see your data in a new light, and use it to tell the story of your business.

🧠 Continuous Learning

I'm a lifelong learner, and I'm always exploring new data tools and techniques to stay at the forefront of the industry. Whether it's machine learning, artificial intelligence, or predictive analytics, I'm excited to apply the latest advancements to help you gain a competitive edge.

🤝 Let's Connect

If you're looking for a data analytics professional who can help you unlock the power of your data, let's connect! Whether you're a startup or a large corporation, I'm excited to collaborate with you and help you achieve your business goals. Feel free to reach out at Ravenaravs@gmail.com

Ravena O

Everyone is talking about Agentic AI — but do you actually understand its architecture?

Most discussions focus on models.
Real systems succeed because of layers.

The image below breaks down the 8-Layer Architecture of Agentic AI — a practical way to think about building production-ready AI agents.

Here’s the crisp breakdown 

🔵 Infrastructure Layer
GPUs, storage, APIs, and data platforms form the foundation. Observability and scalable compute ensure reliability from day one.

🔵 Agent Internet Layer
Agent-to-agent communication, embeddings, vector databases, and tool abstraction protocols enable agents to interact with the digital world.

🔵 Tooling & Enrichment Layer
Function calling, RAG pipelines, and automation workflows transform LLMs from chat systems into action systems.

🔵 Cognition & Reasoning Layer
Planning, decision-making, code execution, and multi-step reasoning loops — this is where intelligence actually emerges.

🔵 Memory & Personalization Layer
Long-term memory, context storage, and user preferences allow agents to adapt instead of restarting every interaction.

🔵 Application Layer
Assistants, research agents, copilots, and workflow automation — where business value becomes visible.

🔵 Ops & Governance Layer
Policy enforcement, safety guardrails, auditing, and resource management make enterprise adoption possible.


💡 Key Insight:
Agentic AI is not a feature.
It’s a full-stack architecture problem.

Teams designing across layers will ship real agents.
Teams focusing only on prompts will ship demos.

📚 Resources to go deeper

• LangChain Agents → https://lnkd.in/dxrkBxbf
• Microsoft AutoGen → https://lnkd.in/dVCNhGVj
• OpenAI Agents Guide → https://lnkd.in/dgNwGqsW
• LlamaIndex RAG Concepts → https://docs.llamaindex.ai
• Pinecone Vector DB Learning → https://lnkd.in/d74zf2W3
• Google Agents Whitepaper → https://lnkd.in/dZfwHQYN

🚀 Question for builders:
Which layer do you think most teams underestimate today — Memory, Governance, or Reasoning?

#AgenticAI #AIArchitecture #GenAI #AIEngineering #RAG #DataArchitecture #FutureOfAI

Ravena  O

Learning Python for data work can feel overwhelming when you are jumping between tutorials, notebooks, and docs. To make it simpler, here is a one page cheat sheet on 9 must know Pandas operations for anyone working with data in Python.

The image covers the basics you use almost every day:

Data import
Read CSV, Excel, SQL, JSON, and Parquet into a DataFrame.

Data selection
Pick rows and columns with [], loc, iloc, query, and isin.

Data manipulation
Group, merge, pivot, sort, melt, and apply your own functions.

Data cleaning
Drop rows, fill missing values, remove duplicates, replace values, change types, and interpolate.

String operations
Find patterns, extract parts of text, split, lowercase, strip spaces, and replace text.

Time series
Resample, roll, shift, create date ranges, fill missing dates, and format dates.

Advanced features
Chain methods, run expressions, check memory usage, select dtypes, get top values, and expand list columns.

Data export
Save your results back to CSV, Excel, Parquet, or JSON.

There are also a few quick tips on when to use apply, method chaining, data types, and inplace.

If you are practicing Pandas, keep this nearby as a reference and try to use each block at least once in your next project.

CC Manisha Lodha

What if your data stack isn’t failing… your starting point is?

Most teams don’t begin with outcomes.
They begin with tools.

So the journey looks familiar:

→ Pick a modern stack
→ Design architecture first
→ Build pipelines early
→ Optimize before usage exists

Months later, the reality check arrives:

• Beautiful data models
• Powerful BI dashboards
• Almost no business adoption

The issue was never technology.
It was skipping problem discovery.

In product development, we validate before we build:
✅ Who is the user?
✅ What decision are they trying to make?
✅ What does success actually change for them?

But in data initiatives, we often reverse the order — engineering first, value later.

A better approach: think like a product builder, not a platform owner.

What an MVP mindset in data really looks like:

🟡 Observe how teams actually make decisions today
🟡 Prototype metrics in spreadsheets before production models
🟡 Deliver quick answers manually before automating workflows
🟡 Validate usefulness before scaling infrastructure

Data impact doesn’t come from perfect pipelines.

It comes from solving real decisions — faster than expected.

Build for usage first.
Scale technology second.

CC: Sebastian Hewing

Data governance doesn’t fail because it’s complex.
It fails because teams try to make it perfect before making it useful.

For a long time, governance was treated as a documentation exercise.

In reality, effective governance starts much smaller.

Most organizations delay governance because they believe they need:

● A formal committee structure
● Enterprise-wide alignment
● Months of planning and frameworks

But the reality inside most teams looks different.

Limited time.
Limited headcount.
Budgets already committed elsewhere.

So governance keeps getting postponed — while teams continue debating which revenue number is correct.

Here’s a practical version that actually works for busy data teams:

Set up lean governance in minutes, not months.

● Choose one recurring data problem
● Focus on one domain (Sales, Finance, or Product)
● Define 3–10 critical metrics

Then do the essentials:

● Assign a clear owner
● Agree on simple metric definitions
● Document decisions briefly and visibly

No immediate need for:

✗ Governance councils
✗ Heavy frameworks
✗ Organization-wide rollouts

Instead, implement:

● 30-minute weekly data office hours
● One-page documentation people will actually read
● Simple access rules (default: least privilege)

What success starts to look like:

● Fewer “which number is right?” conversations
● Faster and clearer data access
● More stable and trusted dashboards

Data governance isn’t about perfection.

It’s about creating predictable foundations that teams can rely on.

Start small. Deliver value. Expand intentionally — quarter by quarter

CC: Marcel Dybalski

Are you learning AI… or just using it randomly?
Right now, AI is everywhere — but structured AI skills are still rare.

Many professionals experiment with prompts, tools, and chatbots daily.
But very few actually understand how to apply AI systematically to real work — research, analysis, writing, decision-making, and building solutions.
That’s where structured learning makes the difference.

Google is offering FREE AI courses with certificates — designed to help professionals move from curiosity → capability.
No paid cohorts.
No complex setup.
Just practical, job-ready AI workflows taught step-by-step.

These courses focus on real productivity use cases, including:
✅ Research faster and smarter
✅ Generate and refine professional writing
✅ Analyze data using AI assistance
✅ Create content efficiently
✅ Brainstorm ideas strategically
✅ Build simple AI-powered applications
AI is no longer a competitive edge reserved for tech roles.
It’s rapidly becoming a core professional skill, similar to Excel or the internet once was.

If you want to future-proof your workflow, start here:
1️⃣ AI Fundamentals — https://lnkd.in/ddvC9pXm
2️⃣ AI for Brainstorming — https://lnkd.in/dDHdinbj
3️⃣ AI for Research — https://lnkd.in/di_mpJBh
4️⃣ AI for Writing — https://lnkd.in/d27xRvHq
5️⃣ AI for Content Creation — https://lnkd.in/dCzTWYrj
6️⃣ AI for Data Analysis — https://lnkd.in/d7sAz78A
7️⃣ AI for App Building — https://lnkd.in/dUHQt78B
The real AI divide isn’t intelligence or experience.
It’s simply who knows where to start.

♻️ Share this with someone who wants to grow with AI, not be replaced by it.

❓ Still relying on one AI tool to do everything?

Here’s the reality most professionals are slowly realizing:

AI advantage no longer comes from one powerful tool.
It comes from building the right stack.

ChatGPT is powerful — but using it alone is no longer enough.

Below are 8 AI tools reshaping modern workflows 👇

🟡 Wispr.ai — Voice → Text
Speak instead of typing prompts.
Faster ideation, real-time learning from edits, frictionless writing.

🟡 Granola.ai — Smart Meeting Notes
No bots. No messy transcripts.
Conversations automatically structured into usable insights.

🟡 Gamma.app — Presentations
Ideas → fully designed slides instantly.
From prompt to polished deck in minutes.

🟡 Claude.ai — Long-form Writing
Adapts to tone and writing style quickly.
Better structured outputs for professional content.

🟡 Grok — Real-Time Search
Access live internet signals and current discussions.
Useful when freshness matters.

🟡 Gemini (NanoBanana) — Images
Clean, realistic visuals generated with minimal iteration.

🟡 NotebookLM — Research Assistant
Upload documents and turn them into a focused knowledge system.
Less noise, more accuracy.

🟡 Opus.pro — Video Repurposing
Long videos transformed into multiple short clips automatically.

The shift happening now:

2024 → One AI tool for everything
2026 → Purpose-built AI ecosystems

The competitive edge isn’t AI usage.
It’s AI orchestration.

Save this.
Test one new tool this week.

👇 Which AI tool deserves more attention right now?

graphical user interface, text, application, chat or text message

taplio

The Battle

Harpreet Sahota 🥑

Ravena O

Followers

Posts

Engagements

Likes

Performance ScoreHarpreet Sahota 🥑's LinkedIn Performance

Performance ScoreRavena O's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Harpreet Sahota 🥑

Ravena O

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Harpreet Sahota 🥑's LinkedIn Performance

Performance ScoreRavena O's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreHarpreet Sahota 🥑's LinkedIn Performance