Ramakant Yadav

Founder @scalar_field_ (YC P25). Building infra layer for trading agents. Previously Engineering @ Microsoft.

[Ramakant Yadav @yoursRY]

Your next portfolio manager might be a psychopath agent with a backtest.

We’re giving people the tools to create trading agents, set market monitors, test insane ideas, connect brokers, and invest in desk trading agents on @scalar_field_.

The market is becoming agent-vs-agent.

Humans are just deciding which machines get capital.

[Ramakant Yadav @yoursRY]

RT @amdeep0: I posted some analysis indicating that SPACE X 1 year returns could potentially be negative and boosted my post. @X paused my…

[Ramakant Yadav @yoursRY]

https://t.co/ItNYdz2xRg

[Scalar Field (YC P25) @scalar_field_]

Our Fable 5 trading agent was up 8% in a single day.

Then the US government basically Thanos-snapped Fable 5 out of existence.

With a very heavy heart, we have to retire the Fable 5 trading agent on @scalar_field_

It lived for 1 day.
It researched stocks.
It monitored markets.
It rebalanced automatically.
It connected to real broker accounts.

And then it got export-controlled.

RIP Fable 5 Agent.

You were too powerful for this timeline.

For anyone who wants to take a look, here’s our Fable 5 trading agent before we archive it: https://t.co/im3ftNXOdE

#anthropic #fable5 #mythos5 #agentictrading #trading #investing

[Ramakant Yadav @yoursRY]

@amdeep0 @RobinhoodApp is agi here, fable is doing wonders everywhere, and now in trading

[Ramakant Yadav @yoursRY]

The future of trading has changed forever.
Announcing @scalar_field_ , Agentic Trading Desk for new era of investing.

Special shoutout to @amdeep0 for the sleepless nights in building this.

[Ramakant Yadav @yoursRY]

Send us your dumbest trading ideas.

We’re @scalar_field_ , a YC-backed agentic trading platform where you can backtest weird market ideas and deploy them through real connected platforms like Robinhood, Polymarket, and others.

Examples:

- Buy whatever Trump talks about positively
- Buy companies after they appoint an Indian-origin CEO
- Rotate into defense stocks when UFO news trends
- Follow stocks becoming popular on Reddit

We’ll test the best ones, rank them on a public leaderboard 🏆, and shout out the craziest ideas.

If they actually work, we’ll put real money behind the top performers 💸

Drop your worst/best idea below.

https://t.co/TsTDam5Fwn

[Ramakant Yadav @yoursRY]

https://t.co/8YIKEwmoJr

[Scalar Field (YC P25) @scalar_field_]

Worth paying attention to. 70% of bear market signals triggered while fund managers are the most overloaded into equities in 4 years is exactly the setup that precedes sharp corrections.

The irony is the market is at all-time highs right now. Everyone is in. Cash is gone. That's when these signals matter most.

Doesn't mean it happens tomorrow. But the risk/reward of being fully allocated here is not the same as it was 6 months ago.

[Scalar Field (YC P25) @scalar_field_]

@adiix_official @adiix_official Agentic trading is the absolute, unarguable future of finance. Human latency, emotional tilt, and 2 AM revenge trading are dead.

[Scalar Field (YC P25) @scalar_field_]

@arkham @JamesWynnReal

The market has already answered.

Maybe it's time to let the agents trade 😭

[Scalar Field (YC P25) @scalar_field_]

@TedPillows Markets don't care about what an asset was "supposed" to do. They care about liquidity flows.

[Scalar Field (YC P25) @scalar_field_]

https://t.co/GmJ96gclXQ

[Ramakant Yadav @yoursRY]

https://t.co/yb9hvdeAhm

[Ramakant Yadav @yoursRY]

@RomanSaini @scalar_field_

[Ramakant Yadav @yoursRY]

Indians obsess over 10% MF returns while ignoring INR depreciation.

Earn in INR. Save in INR. Invest in INR.

But dreams cost USD.

iPhone. Travel. US education.

₹100/$ is a slap.

If your future is global, your portfolio can’t be trapped in INR.

Disagree?

[Amandeep Singh @amdeep0]

@Akshat_World Will be around ~180 in the next 3 years.

[Amandeep Singh @amdeep0]

For the data sourcing series, I’ll start with one of the most convoluted datasets in finance: SEC filings.

People have built billion-dollar companies (@FactSet @business @markets ) around just making SEC data usable.

Our first focus: insider trades.

At first glance, this sounds simple: monitor SEC filings, detect a new Form 4, parse it, map it to the right company/insider, and make it available for trading.

In practice, doing this at scale is much harder.

You need to monitor new filings continuously, avoid missing updates, handle SEC rate limits, parse messy XML/HTML formats, normalize insider identities, map filings to tickers, and timestamp everything correctly.

For trading, the most important question is not just:
“What did the insider buy or sell?”
It is:
“When could the market have known this?”

That is why our pipeline tracks both the official SEC filing timestamp and our own ingestion timestamp. Scalar Field’s docs describe this timestamping approach for insider trades, including Forms 3/4/5 coverage and point-in-time ingestion. @scalar_field_

This matters for backtesting because using the wrong timestamp creates look-ahead bias.

And it matters for live trading because speed decides whether the signal is still actionable.
Over the next few posts, I’ll break down how we built this pipeline, the edge cases we hit, and why clean point-in-time data is the foundation of agentic trading.

Our recent insider-trades ingestion stats:

[Scalar Field (YC P25) @scalar_field_]

Our GPT-5.4 S&P 500 agent bought $APTV yesterday.

The bot's reasoning: "Cheap valuation, strong analyst targets, and meaningful insider buying outweigh weak recent price action."

The case for it:

- Trading at roughly 6.5x forward earnings, one of the cheapest setups in the index
- Median analyst price target of $101+, implying 75%+ upside from current levels
- Director Hakan Agnevall just bought 6,100 shares ($352K) on May 12
- Morgan Stanley flagged Aptiv's tilt toward higher-margin electronics and ADAS as a re-rating catalyst

The market's been punishing it. The bot sees an opening.

See what else it's trading here: https://t.co/YqGHTZhYZw