Ajeet Verma 🇮🇳

Lead AI Engineer at FPT Software | Data Scientist

[Elon Musk @elonmusk]

Robot fights are fun 😂

[Ajeet Verma 🇮🇳 @ajeet214]

https://t.co/cE2OocdXb2

[LangChain @LangChain]

✨ The Definitive Guide to Testing LLM Applications by LangChain ✨

Reviewing agent responses can be a time-consuming and daunting process, from defining criteria for style and accuracy, to spotting new regressions.

After partnering with hundreds of companies to enhance their agents, we've put together a comprehensive guide of best practices for testing throughout the development lifecycle.

In this guide, you'll learn how to:
• Tips for testing across the product lifecycle
• Methods for building a dataset & defining testing metrics
• Templates for evaluating agents, with visual examples
... and much more!

👉 Get the guide here: https://t.co/4d8j3ZpnqY

[Ajeet Verma 🇮🇳 @ajeet214]

True

[Top 1% Men @dtop1percentmen]

This video changed my way of thinking

[Ajeet Verma 🇮🇳 @ajeet214]

@elonmusk As we say, history repeats itself. The balance of power is shifting back to where it used to be once upon a time in 17th-18th century when India + China alone accounted for 48-50% of world’s GDP

[Harrison Chase @hwchase17]

LangChain vs langgraph vs deepagents

When to use each one

[Ajeet Verma 🇮🇳 @ajeet214]

https://t.co/1gNYHuwReE

[Ajeet Verma 🇮🇳 @ajeet214]

https://t.co/UMGlWE4YkF

[Akshay 🚀 @akshay_pachaar]

Everyone building AI agents keeps making the same database mistake!

They give agents direct read/write access to their production database, thinking that rate limits and permissions will keep things safe.

Here's what actually happens:

The Agent runs 50 parallel queries to optimize something.

Each query seems reasonable in isolation. Together, they create a cascade of locks, updates, and resource consumption that brings down your entire system.

Or worse: The agent decides to test something, and to do that:

- It creates a new index.
- Drops it.
- Creates another.
- Tests a migration.
- Rolls it back.
- Test another approach.

All on your live data.

Even the best agents are experimental by nature. They try things, fail, learn, and try again (think about how many times Cursor says, "let me try an alternative approach").

This experimental nature is exactly what makes them powerful.

But that's also exactly what makes them terrifying when they touch production systems because traditional databases were never designed for this behavior pattern.

Postgres was built for humans who carefully plan queries, test in staging, and deploy with caution.

But Agents don't work like that.

They explore, branch, parallelize, and iterate at machine speed.

So how do you give agents the database access they need without the existential dread?

You need databases that can fork instantly.

Imagine this: Your agent needs to test a schema migration. Instead of touching production, it spins up a perfect copy of your entire database in seconds.

The agent can now destroy this fork completely. Try 20 different indexes. Run performance benchmarks. Test edge cases. When it finds the optimal solution, it reports back. When it doesn't, you just delete the fork.

This exact solution is actually implemented in the newly released Agentic Postgres by @TigerDatabase, which you can use for free.

Tiger Data built this after watching agents struggle with traditional databases for years, and they worked with me on this post to share what they did.

But the innovation in Agentic Postgres does not just stop at forking.

It has a built-in MCP server that teaches Agents how to actually think in Postgres.

Instead of generic SQL that mostly works, the MCP server aggregates 35 years of Postgres knowledge that Agents can reason with.

It also has a native hybrid search (vector + BM25) so agents can retrieve and learn from data without external tools.

And they added memory APIs for persistent context across sessions.

In the video below, I used it with Claude Code, and the difference is evident.

You can try it yourself here: https://t.co/EPSAS8coQr

[Akshay 🚀 @akshay_pachaar]

Google dropped another banger!

They just released a comprehensive white-paper on AgentOps - the missing piece between building AI agents and actually shipping them to production.

Here's the reality:

Building an AI agent takes minutes. Making it production-ready? That's where 80% of the real work begins.

Google's "Prototype to Production" guide tackles this exact problem.

The framework has three core pillars:

1. Evaluation-Gated Deployment:

No agent reaches users without passing tests. Build a "golden dataset" that validates behavior, not just functionality. This catches what unit tests miss - agents choosing wrong tools or hallucinating responses.

2. Automated CI/CD for Agents:

Test in stages: pre-merge checks for fast feedback, staging for load testing, then gated production. Version everything: prompts, tools, configs, evaluation datasets.

3. Observe → Act → Evolve Loop

Production isn't the finish line. Monitor through logs, traces, and metrics. Act with circuit breakers and human escalation. Evolve by turning production failures into test cases.

The best part? They released the Agent Starter Pack - a template with CI/CD, Terraform deployment, and built-in observability. Helps you spin up an evaluation pipeline in minutes.

The guide also talks about the two major protocols and how they can work together.

↳ MCP for tool integration
↳ A2A for agent collaboration

If you're shipping agents to production, you should read this.

I've shared the full white-paper in the next tweet!

[Avi Chawla @_avichawla]

LangGraph.
CrewAI.
Agno.

Which one to pick?

The good news is that this will not matter soon!

Finally, we have a full picture of how the industry is solving this with just three open protocols that work across ALL frameworks.

It's not about picking the best framework.

Instead, it's about understanding how protocols create interoperability.

The Agent Protocol Landscape shows how three complementary protocols are creating a universal language for Agents:

> AG-UI (Agent-User Interaction):
- The bi-directional connection between agentic backends and frontends.
- This is how agents become truly interactive inside your apps, not just as chatbots, but collaborative co-workers.

> MCP (Model Context Protocol):
- The standard for how agents connect to tools, data, and workflows.

> A2A (Agent-to-Agent):
- The protocol for multi-agent coordination.
- How agents delegate tasks and share intent across systems.

These aren't competing standards. They're layers of the same stack and have handshakes with each other.

So instead of building point-to-point integrations, you build to protocols.

Moreover, you can integrate LangGraph, CrewAI, or Agno into the same frontend, without rewriting your UI logic. These protocols let everything work together.

For instance:

- Your LangGraph agent pulls data via MCP.
- It delegates analysis to a CrewAI agent via A2A.
- Results stream to your React app via AG-UI.
- Users see real-time collaboration in your interface.

This way, you can focus on building agent capabilities instead of integration mechanics. The protocols handle interoperability automatically.

CopilotKit unifies this entire stack into one framework so you can build "Cursor for X" style apps without implementing each protocol from scratch.

It gives you all three protocols, generative UI support, and production-ready infrastructure in one framework.

I have shared this playbook in the replies!

It breaks down handshakes, misconceptions, and real examples and shows exactly how to start building.

[Ajeet Verma 🇮🇳 @ajeet214]

𝐓 – 𝐓𝐞𝐬𝐭 (𝐨𝐫 𝐓𝐚𝐫𝐠𝐞𝐭):

Verify or specify the expected output format, evaluation criteria, or how the response will be checked.

𝐄𝐱𝐚𝐦𝐩𝐥𝐞: “Output must be in JSON format with fields: risks, outlook, recommendations.”

[Ajeet Verma 🇮🇳 @ajeet214]

What is the 𝐒𝐀𝐋𝐓 framework in prompt engineering?

The SALT framework in prompt engineering is a structured approach used to design, refine, and evaluate prompts for large language models (LLMs).

#PromptEngineering #LLMs #genai #ai

[Ajeet Verma 🇮🇳 @ajeet214]

𝐋 – 𝐋𝐢𝐦𝐢𝐭𝐬:

Add boundaries, rules, or constraints to control the output.

𝐄𝐱𝐚𝐦𝐩𝐥𝐞: “Keep the summary under 200 words, use bullet points, and avoid technical jargon.”

[Ajeet Verma 🇮🇳 @ajeet214]

--𝐬𝐭𝐚𝐭, quick overview of how large each commit is & what files changed

[Ajeet Verma 🇮🇳 @ajeet214]

Most devs only use 𝐠𝐢𝐭 𝐥𝐨𝐠 to see history.
But 𝐠𝐢𝐭 𝐥𝐨𝐠 is actually a time machine you’re not using properly.
Here’s how to unlock it 👇

#git #github #gitlog #devtools #versioncontrol

[Ajeet Verma 🇮🇳 @ajeet214]

--𝐠𝐫𝐞𝐩, find commits by topic / bug ID / Jira ticket number etc.