Prabhash Varma

Software Engineer

[Lokesh Nara @naralokesh]

This is how fake news is manufactured - and then amplified by senior Congress leaders without even bothering to verify the facts.

There was no disruption of the Finance Minister’s speech by local youth demanding jobs. The claim being circulated is completely false. A normal interaction with an enthusiastic gathering is being deliberately clipped, twisted and given a fabricated narrative.

Shri @Jairam_Ramesh, surely a senior leader of your experience should verify facts before tweeting.

Deliberate misinformation is not on. Delete the lie.

@nsitharaman @ncbn

[​𝐥𝐲𝐫𝐚 @sunnkssdseraph]

People who spend long periods of time at home and slowly lose the desire to go outside often think they're just resting. But psychologically, many of them are experiencing something called mental rumination. It's a state where the brain starts slowing down while the number of thoughts inside keeps increasing and the motivation to do anything slowly disappears. The mind gets stuck in the past, worries about the future and burns itself out. This isn't laziness, it's a nervous system stuck in survival mode. And the longer it lasts, the more distant normal life begins to feel. And the difficult part is that you usually can't escape it through one massive burst of motivation. And sometimes,recovery starts with just one small step.

[Prabhash Varma @Prabhashvarma]

@Srav1911 @rohithdoesdev 🤣🤣🤣

[Prabhash Varma @Prabhashvarma]

@oldboy2011 Congratulations 🎉

[Palo Alto Networks @PaloAltoNtwks]

Deploying coding assistants and agents without governance or guardrails can be dangerous.

Prisma AIRS AI Gateway delivers a centralized control plane to discover AI usage, govern interactions, and secure agent identities at runtime.

Take control of your AI ecosystem. https://t.co/SIodBvqHgv

[ClaudeDevs @ClaudeDevs]

https://t.co/v694m8Eaj6

[Aadhithya D @AadhithyaD_]

Launching Raaga V1 by Vakyam AI.

Natural, expressive speech for Indian languages.

Frontier-level quality at ₹0.75 per 1K characters (~per min).

Try it now on our playground.

[Prabhash Varma @Prabhashvarma]

Didn't expect the GPT-5.5 drop to out-hype Opus 4.7 but here we are 😂

[Ayoub @ayoubAIdev]

We are all waiting for gpt 5.5 are we?

[Portkey @PortkeyAI]

Agent Gateway is live in beta

Register any agent, get a governed Portkey endpoint, and every action flows through a single control plane.

[siddharth @sambharia]

WE JUST SHIPPED PORTKEY'S AGENT GATEWAY 🚀

when i joined portkey in 2024, teams were building simple gpt wrappers. copy a prompt, call an api, ship a chatbot. we solved for that world — a unified api that let you connect multiple llms reliably, with full visibility into every call.

that world doesn't exist anymore. enterprises today aren't shipping chatbots. they're shipping agents. agents that run autonomously, chain tool calls, consume mcp servers, make decisions, and operate over longer timescales with actual real-world consequences.

and every single team that ships an agent hits the same wall. fast.
who is allowed to call this agent? what did it cost to run? what did it actually do? which tools did it touch? and when something goes wrong — who is responsible?

right now every team patches this together themselves. duct tape, custom logging, vibes-based access control. it's not scalable and everyone knows it.

so we built the infrastructure that solves this properly.

Portkey's agent gatewaw is the production layer between your agents and the world. built on the same infra that processes 1T+ tokens every day across thousands of enterprise teams.
here's what ships today:

- full governance and access control for every agent and tool call
- end-to-end observability across entire agent runs
- agent registry to manage, version, and audit your agents
- guardrails so your agents don't go off the rails in prod
and a lot more

we're doing a live webinar on April 24th where we demo everything we've been building.

if you're shipping agents in production (or planning to), you don't want to miss this.
link below 👇

[Portkey @PortkeyAI]

🚀 Claude Opus 4.7 is live on Portkey!

@AnthropicAI's most capable Opus model showed improvements in complex and long running coding tasks.

Try it today!

[Ben Lang @benln]

Cursor team is coming to India:

• Bangalore - 4/25
• Chennai - 5/1
• Mumbai - 5/9

Sign up for the meetups on Luma

[siddharth @sambharia]

https://t.co/3iSPkL7Lne

[Prabhash Varma @Prabhashvarma]

@RoxanaLimban Hey! Let's connect

[Prabhash Varma @Prabhashvarma]

@sync_ise Let's connect

[Prabhash Varma @Prabhashvarma]

@isha_singh06 Hi, Let's connect

[Prabhash Varma @Prabhashvarma]

@nxrxnxn This post got me stressed 😂😂

[Rohit Agarwal @jumbld]

Most SaaS companies are on their way to becoming glorified databases.

Said this out loud on the Neon Show with @siddharthaa7 and then had to explain myself.

We got into why enterprises are spending MORE on AI models, not less, and much more, including if Portkey has any moat or not!

Was a fun convo

[The Neon Show @TheNeonShoww]

Over 500 billion AI tokens pass through Portkey every single day.

Every team shipping AI in production hits the same wall. The prototype works, but at scale the system has to juggle multiple models, manage token costs that compound fast, absorb latency spikes, and hold together on infrastructure that was never built for AI workloads. Most companies underestimate this layer until it becomes a problem.

That is the problem Rohit Agarwal is solving with Portkey, an AI gateway that sits between applications and the models, whether that’s GPT-4, Claude, or Gemini.

With 24,000 companies routing their AI through Portkey, Rohit sits on ground-level data on how AI is actually being used in production. Which models enterprises are betting on. Where costs are quietly climbing. How usage patterns shift as companies move from pilots to real products.

When AI spend surpasses cloud spend, and Rohit believes it will, the infrastructure layer sitting underneath it becomes one of the most important bets in enterprise tech.

This episode explores what it takes to run AI systems at that scale.

@PortkeyAI | @jumbld

Catch up the full conversation here: https://t.co/Rcc3N6O9Fv