Sign in

Maria Khalusova

@mariak.bsky.social
3.2K followers 137 following 187 posts

Always growing, she/her, RAG builder, LLM whisperer, tech generalist

PostsRepliesMedia
Maria Khalusova @mariak.bsky.social · 19/12/2025
Really looking forward to food-induced coma naps
010
Maria Khalusova @mariak.bsky.social · 15/12/2025
What's going to be different for you in 2026?
100
Maria Khalusova @mariak.bsky.social · 01/12/2025
Just entered my next decade and I think it’ll be the best one yet.
120
Maria Khalusova @mariak.bsky.social · 24/09/2025
Once again, I completely forgot that I have this account. Oops
010
Maria Khalusova @mariak.bsky.social · 13/06/2025
Things move fast in AI. Every week brings new models, new capabilities, or new ideas to chase. It’s exciting, but also easy to get swept up in the pace and forget to pause, to touch grass, to zoom out and see the bigger picture. 🧵
120
Maria Khalusova @mariak.bsky.social · 12/06/2025
Asking “What is the best chunk size for RAG?” without any additional context is like asking, “What’s the best thing to wear?” Wear where? What’s the weather like? What size are you? Are you going to a wedding or hiking a trail? There’s no single answer that works for every situation. 🧵
100
Maria Khalusova @mariak.bsky.social · 09/06/2025
Do the thing that you will regret not doing when you're old.
020
Maria Khalusova @mariak.bsky.social · 30/05/2025
I went to check what new courses deeplearning.ai has, and was pleasantly surprised to see that the short course Marc Sun, Younes Belkada, and I have built over a year ago is still featured as one of the Top Rated courses 😍
060
Maria Khalusova @mariak.bsky.social · 20/05/2025
I'm taking this whole developer becoming a farmer dream way too far, am I?
110
Maria Khalusova @mariak.bsky.social · 15/05/2025
If you've been prioritizing urgent work, make sure to prioritize important work.
020
Maria Khalusova @mariak.bsky.social · 12/05/2025
How anyone can like peanut butter is beyond me.
100
Maria Khalusova @mariak.bsky.social · 09/05/2025
Similar ≠ relevant
010
Maria Khalusova @mariak.bsky.social · 07/05/2025
Nothing starts a Wednesday morning quite like your dog getting sprayed by a skunk 🤢
120
Maria Khalusova @mariak.bsky.social · 02/05/2025
I have some epic plans for this summer and none of you’ll be able to guess what they are.
100
Maria Khalusova @mariak.bsky.social · 01/05/2025
I'm starting a series of blog posts on RAG beyond the basic set up. In the first part, we're setting the stage. Why naive RAG is not enough, and how a lot of the issues can be traced back to data processing choices. Part 1: unstructured.io/blog/level-u...
unstructured.io
Level Up Your GenAI Apps: RAG Beyond the Basics – Unstructured
Learn why naive implementations fall short—and how smarter data preprocessing lays the foundation for reliable, high-performance RAG.
150
Maria Khalusova @mariak.bsky.social · 30/04/2025
What you're not changing, you're choosing. This is a gentle reminder for the next time you're prioritizing a cool new shiny thing over building the foundation or addressing tech debt.
010
Maria Khalusova @mariak.bsky.social · 29/04/2025
Word of the day seems to be "sycophantic". Thanks AI community for increasing my vocabulary :)
000
Maria Khalusova @mariak.bsky.social · 28/04/2025
32 pages - is it still a blog or do I call it a book now?
030
Maria Khalusova @mariak.bsky.social · 25/04/2025
I don’t like the term “AI-assisted coding”. AI is just another tool in the box. We've had code completion and refactoring features in IDEs for a hot minute now, do we call it “IDE-assisted coding”? We just use these things to make our lives easier. 🧵
120
Maria Khalusova @mariak.bsky.social · 24/04/2025
Apropos of nothing, here's a tip. Some embedding models are robust against typos, and others are sensitive to them. That's a whole rabbit hole on its own 🤓, but long story short - this will matter if your users are fat-fingering the queries.
110
Maria Khalusova @mariak.bsky.social · 24/04/2025
Today I learned that Postgres doesn't delete columns. Not from the catalog and not from data files. Deleted columns are only "soft deleted". But they still exist. And now I wonder if this is why the logo is an elephant. As in, it never forgets.
161
Maria Khalusova @mariak.bsky.social · 15/04/2025
I think I pissed off Live Demo gods at some point, they were not on my side today.
110
Maria Khalusova @mariak.bsky.social · 14/04/2025
I may need to travel to the US soon, and this genuinely worries me.
110
Maria Khalusova @mariak.bsky.social · 11/04/2025
I'm looking for incredible people to join my team! If you're excited about AI and know that great AI starts with great data - If you love building, learning, and helping others do the same - DM me. Let's talk. And if this sounds like someone you know, feel free to RT!
111
Maria Khalusova @mariak.bsky.social · 11/04/2025
We use RAG to look up context for an LLM, next step should be looking up available MCP Servers and their tools. We need a middle layer between the LLM and servers, to look up available tools based on the request instead of having all of that info blow up the context window.
120
Maria Khalusova @mariak.bsky.social · 10/04/2025
April appears to be a good month for hitting milestones.
040
Maria Khalusova @mariak.bsky.social · 03/04/2025
Years of breaking down complex concepts to a broad audience turn out to be rather useful in vibe coding.
040
Maria Khalusova @mariak.bsky.social · 02/04/2025
Figuring out Jira automations is not something I had envisioned for myself, but here we are 🤣
000
Maria Khalusova @mariak.bsky.social · 02/04/2025
Lol. I just saw on LinkedIn a photo of a person giving a talk (from a meetup this week), and on a screen behind them there was a slide, word for word exactly the same as I have in my deck from a talk I gave in February. I'm choosing to be flattered and amused rather than pissed off.
390
Maria Khalusova @mariak.bsky.social · 28/03/2025
Even in the AI age, the hardest things are still naming, and auth. 🫠
030
Maria Khalusova @mariak.bsky.social · 26/03/2025
Day #3. It's so on! 👩‍💻
000
Maria Khalusova @mariak.bsky.social · 25/03/2025
Day #2 of Unstructured MCP Hackathon! Don't miss it!
021
Maria Khalusova @mariak.bsky.social · 24/03/2025
Want to learn how to build MCPs? The best way to learn is by building! So, Unstructured just launched a virtual MCP hackathon: 🔧 Build an MCP server with our API 📝 Create a tutorial to share your work 💰 Win one of three $500 prizes Starts today! Rules here: unstructured.io/blog/unstruc...
unstructured.io
Unstructured MCP Virtual Hackathon: Build, Share, and Win! – Unstructured
Join our first-ever Virtual Hackathon, build MCP servers with Unstructured API under the hood, and win prizes!
030
Maria Khalusova @mariak.bsky.social · 20/03/2025
Thoughts on MCP. First, the cool thing about the MCP is that it’s a universal adapter to all kinds of MCP hosts that have an MCP client - Claude Desktop, Cursor, Windsurf, etc. You only need to build one MCP server, and it becomes a tool for all the LLM-backed clients you do NOT develop. (continued)
150
Maria Khalusova @mariak.bsky.social · 11/03/2025
Some progress on my MCP tinkering. I've successfully created an ephemeral workflow to process given file with Unstructured Platform via the API and clean up everything after, now I just need to turn it into an MCP tool and see how I can use it from Claude Desktop. Let's go!
110
Maria Khalusova @mariak.bsky.social · 10/03/2025
I'm starting to look into MCP. Thinking of building a local MCP server over the next few days to extend Claude Desktop's range of files it can take in by processing files with Unstructured Platform.
020
Maria Khalusova @mariak.bsky.social · 27/02/2025
I have massively enjoyed presenting at ConFoo Montreal today to a packed room of RAG enthusiasts. Thanks to all who came to my talk! As promised, you can find the slides with a list of additional resources at the end here - github.com/confooca/202...
github.com
080
Maria Khalusova @mariak.bsky.social · 25/02/2025
Do you ever sign up for webinars knowing full well that you will not attend, but instead will watch the recording they send afterwards at 2X speed? Or is it just me?
1100
Maria Khalusova @mariak.bsky.social · 21/02/2025
My most important plans for 2025 have nothing to do with AI.
030
Maria Khalusova @mariak.bsky.social · 04/02/2025
Unpopular opinion: good food is supposed to cost more. It's only cheap at supermarkets because it's dowsed in pesticides, harvested by people who are underpaid, and grown by destroying the soil with industrial fertilizers.
070
Maria Khalusova @mariak.bsky.social · 03/02/2025
So many tariff experts all of a sudden, almost as many as AI experts.
150
Maria Khalusova @mariak.bsky.social · 31/01/2025
Judging by the lack of comments to this tweet, y'all didn't read anything good last year.
310
Maria Khalusova @mariak.bsky.social · 30/01/2025
What fiction books did you read in 2024 and would recommend? And why?
421
Maria Khalusova @mariak.bsky.social · 29/01/2025
"Facility", right.
011
Maria Khalusova @mariak.bsky.social · 29/01/2025
I said it before, and I'll say it again. LLMs are easy to swap. Data prep and integration, orchestration and good retrieval, evals and monitoring is where it's at.
190
Maria Khalusova @mariak.bsky.social · 27/01/2025
Ever since 2020, it has been a constant balancing act between two opposing forces: the desire to stay informed and the need to remain mentally healthy 🫠
110
Maria Khalusova @mariak.bsky.social · 23/01/2025
Oh no, I have to write the emails myself now :D
041
Maria Khalusova @mariak.bsky.social · 22/01/2025
I would love to touch grass, but it's under a thick layer of snow for the foreseeable future.
120
Maria Khalusova @mariak.bsky.social · 17/01/2025
Forgot a column in my schema when creating a table. This is so convenient! It's the little things sometimes. Thanks PyCharm!
050
Maria Khalusova @mariak.bsky.social · 17/01/2025
Folks at Hugging Face recently released a new lightweight library for building agents called `smolagents`. I've created a notebook that illustrates how to build Agentic RAG with it on a bunch of PDF reports, and how it performs compared to Vanilla RAG: colab.research.google.com/drive/1hG3dP...
colab.research.google.com
Google Colab
160