Nandeshwar Gupta

Senior Machine Learning Engineer @ZupeeOfficial || Working on recommender systems and Generative AI

[Nandeshwar Gupta @nandeshwar_g]

@TataMotors @groupcxm Customer feedback also doesn't appear to be genuinely collected at Tata service centres - it often seems to be filled by staff themselves.

The Harrier is a good product, but the service experience has been so poor that I genuinely don't want to deal with Tata in the future.

3/3

[Nandeshwar Gupta @nandeshwar_g]

@TataMotors @groupcxm Extremely disappointed with the service experience for my Tata Harrier at Autovikas Tata, Faridabad Sector 5.

Sent the car in because a warning symbol was appearing on the odometer. Got it back with ANOTHER engine warning light.

1st message of 3

[Nandeshwar Gupta @nandeshwar_g]

@TataMotors @groupcxm The car was returned extremely dirty, with fender-to-fender scratches on one side and blue paint marks on another.

This is not an isolated experience. Poor communication, staff rarely answering calls, and very poor management at the service centre.

2nd message of 3

[Nandeshwar Gupta @nandeshwar_g]

31st March 2026

[Nandeshwar Gupta @nandeshwar_g]

Atal tunnel 31st March
#manali #ataltunnel #sissu #atal #tunnel

[Santiago @svpino]

Here is what you need to know about Agent Protocols in 2025:

• MCP (Model Context Protocol) → defines how agents access tools, data, and workflows. It's the standard behind "agents with tool use."

• A2A (Agent-to-Agent Protocol) → defines how agents coordinate, delegate, and share intent across systems. It's how multi-agent collaboration happens.

• AG-UI (Agent–User Interaction) → defines how humans and agents collaborate inside real apps. Enables real-time, multimodal, bi-directional interaction between frontends and agentic backends.

In summary:

• MCP connects agents to tools and data
• A2A connects agents to other agents
• AG-UI connects agentic apps to agentic backends

The three protocols work together. They help us build reusable, composable, open, interoperable, and transparent applications.

I'm sharing a link to a PDF, "The Agent Protocol Landscape," from the @CopilotKit team.

This document has everything you need to understand about these protocols, how they work, and how you can take advantage of them.

CopilotKit is an Agentic Application Framework that sits above these protocols. ← This is what you need to build agentic apps that use any/all of these protocols.

It essentially allows you to build a "Cursor for [anything]"-type application, where it handles integrating MCP and A2A, and utilizes AG-UI to transmit protocol messages to the front-end.

Here is the link to the PDF: https://t.co/qAd5Ij30GA

[Nandeshwar Gupta @nandeshwar_g]

@Traffic_hqrs @JmuKmrPolice @JKTransportDept @OfficeOfLGJandK Stuck at same place in jam near Chanderkote for 5 hours towards Jammu

[Nandeshwar Gupta @nandeshwar_g]

@championswimmer Using obsidian for few years now, for big notes. Good sync only available for apple devices. Crazy customisable, good text editor and shortcuts. Don’t use graph much though.

[Nandeshwar Gupta @nandeshwar_g]

@Al_Grigor EMR + Sagemaker

[Radek Osmulski @radekosmulski]

To be in ML is to learn.

[Manthan Gupta @manthanguptaa]

The 100cr salary of @zerodha founders is not eye-opening to me, but the tech and how they operate on such a large scale is to me.

How do they optimize their database with close to 20TB of data?

A 🧵

[Suzana Ilić @suzatweet]

Pretty cool: @DeepMind released the '21 DeepMind x
UCL Reinforcement Learning Lecture Series – comprehensive intro to modern RL. Anyone up for study sessions @__MLT__? 📚

https://t.co/8XT08min58

[Sebastian Raschka @rasbt]

"Introduction to Deep Learning -- 170 Video Lectures from 𝐀daptive Linear Neurons to 𝐙ero-shot Classification with Transformers"
Just organized all DL-related videos I recorded in 2021. Hoping it might be useful for one or the other person out there:
https://t.co/8FhMfL6v3N

[Nandeshwar Gupta @nandeshwar_g]

Day 9 - Linear Regression from scratch using NumPy.💯

I cover Linear regression implementation from scratch using NumPy and Scikit Learn both. 🚀

Day 9 blog - https://t.co/npLPLVjG8D

#deeplearning #artificialintelligence #machinelearning #python #100daysofdeeplearning

[Nandeshwar Gupta @nandeshwar_g]

Day 0 Announcement, Tomorrow is the very first day of 100 Days Of Deep Learning 💯.

Day 0 blog - https://t.co/SMZJaj3CYX
Subscribe to the newsletter - https://t.co/zNclEtH8PI

#100daysofdeeplearning #100daysofdatascience #deeplearning #artificialintelligence #machinelearning

[Nandeshwar Gupta @nandeshwar_g]

I have set myself on a path where I will take 100 Days Of Deep Learning covering all basic and advanced topics. It will be covering theory, maths, implementation, reading research papers and projects.

I will be publishing these on https://t.co/B76SDPaGR0

#100daysofdeeplearning

[Nandeshwar Gupta @nandeshwar_g]

Day 8 - What is Linear Regression with Derivation. 💯

Introduction to Linear Regression and Derivation of Ordinary Least Square method.🚀

Day 8 blog - https://t.co/wPy02LkduJ

#deeplearning #artificialintelligence #machinelearning #python #100daysofdeeplearning #AI

[Nandeshwar Gupta @nandeshwar_g]

@MLsummaries @readwiseio save thread