Shrinath Suresh

GenAI Engineer, Python enthusiast & content creator. I write about end-to-end use cases. Connect:

[Shrinath Suresh @shrinathsuresh]

Happy to host Gowtham Balachandhiran in AI Paatshal channel. Unscripted talk on LLM inference. Video Link - https://t.co/GqFwXW2CsR

[Shrinath Suresh @shrinathsuresh]

Happy to share the third video in the series - Deploying @superlinked Inference Engine (SIE) on GPU - Embeddings and zero shot extraction https://t.co/Pq4mBR5eUt

[Shrinath Suresh @shrinathsuresh]

Creozen is hosting a free workshop this weekend — sign up if you’re looking to get into agents and MCP!

https://t.co/Wk664To4L6

[Shrinath Suresh @shrinathsuresh]

Track your data using MLflow https://t.co/1xX26UA2HI

[Shrinath Suresh @shrinathsuresh]

Prompt Engineering techniques https://t.co/fvrZQ2sm8T via @YouTube

[Shrinath Suresh @shrinathsuresh]

Deploy open source models - cheap and Fast https://t.co/AzpltjuInw via @YouTube

[Shrinath Suresh @shrinathsuresh]

Don't waste money on API-based models for single-task use cases. Instead, fine-tune a model and run it locally for cost-effective experimentation. Need model suggestions?

https://t.co/zNZlDFXMur

#deployment #ai #llm #mlops #machinelearning

[Shrinath Suresh @shrinathsuresh]

Batch Profess with @Doubleword_ for Lower Inference Cost https://t.co/M9cHJZLFzE

#artificialintelligence #llm #inference #mlops #deployment

[Shrinath Suresh @shrinathsuresh]

Server vs serverless inference https://t.co/QFRk6HK4yU via @YouTube

[Shrinath Suresh @shrinathsuresh]

Prompt for reading research papers https://t.co/Kd6bd1vcZ0 via @YouTube

[Shrinath Suresh @shrinathsuresh]

One prompt to master any concept https://t.co/ODiEJvAATe

#prompt #promptengineering #ai #learning #notebooklm

[Shrinath Suresh @shrinathsuresh]

Speech Generation https://t.co/KGv7zlvR83 via @YouTube

[Shrinath Suresh @shrinathsuresh]

Still writing long prompts? Optimize them with Al! Using chat-based LLMs can make prompts generic. Instead, use prompt optimization techniques for better results. Use @MLflow dspy prompt optimizer to trailer prompt to your dataset.

https://t.co/ycKSzrVWfy

[Shrinath Suresh @shrinathsuresh]

Invoice extraction is a complex task that presents interesting challenges for both students and professionals. While it appears simple, ensuring good accuracy at scale is difficult due to multiple reasons.

https://t.co/YAZP0JSrrO

[Shrinath Suresh @shrinathsuresh]

Batch inference for open source llms https://t.co/n2d2Q5CzMx

#ArtificialIntelligence #llm #OpenSource

[Shrinath Suresh @shrinathsuresh]

🚀 Starting a new mini-series on AI Model Deployment using Serverless Architecture

In this video, I explain the difference between server-based and serverless deployment using a simple analogy.

Video: https://t.co/tCWHjHXapX

#ArtificialInteligence #MLOps

[MLflow @MLflow]

Stop treating your training data like a "one-off" CSV. 🛑

We version code. We version models. Why are we still losing track of the exact data used in experiments? 🤔

If you can’t point to the specific slice of data used, your experiment isn't reproducible—it’s just a lucky run. 🍀

In the latest episode of "Getting Started with MLflow," @shrinathsuresh breaks down how MLflow Dataset Tracking closes the gap:

🔹 Log datasets from pandas, Spark, & Hugging Face
🔹 Auto-capture schemas to track changes
🔹 Maintain clear version history as data evolves

Data changes are just as critical as model changes. Stop the guesswork and start tracking.

📺 Watch the full guide here: https://t.co/yKwKvPYCAs

#MLOps #MLflow #DataScience #MachineLearning #Reproducibility #OpenSource

[Shrinath Suresh @shrinathsuresh]

Happy to share my next video, where I give a quick tour of @Doubleword_, a platform designed for running open-source models with both Playground and API-based batch scheduling options.
🎥 Full walkthrough in the video - https://t.co/q2yLuqenDQ

[Shrinath Suresh @shrinathsuresh]

I’ve curated a set of @MLflow features that make LLM observability simpler and more effective. If you’re working on GenAI workflows, this deep dive might be helpful—read the full Medium post here:
https://t.co/4r5bU4utvJ

[MLflow @MLflow]

🚀 𝗚𝗲𝘁𝘁𝗶𝗻𝗴 𝗦𝘁𝗮𝗿𝘁𝗲𝗱 𝘄𝗶𝘁𝗵 𝗠𝗟𝗳𝗹𝗼𝘄: 𝗟𝗼𝗴𝗴𝗶𝗻𝗴 𝗚𝗲𝗻𝗔𝗜 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝘀 𝗖𝗼𝗱𝗲

Why log GenAI models as code? It's essential for ensuring versioning and reproducibility of API-based models, providing clear traceability for the model's logic and dependencies.

In this video with @shrinathsuresh, MLflow Ambassador, learn how to log GenAI models as code directly from Python into MLflow Tracking.

🎥 𝗪𝗮𝘁𝗰𝗵 𝗻𝗼𝘄: https://t.co/ZlMpyVpsZC

#opensource #oss #mlflow #genai #mlops