Company profile · 4 min read
Scalar Field: an agentic trading desk
Scalar Field connects market research, backtesting, event-driven execution and portfolio state for agents that operate across supported venues.
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
Published · Updated
What it does
Scalar Field is building an agentic trading desk: one place to research an investment idea, backtest a strategy, monitor events and turn it into a live portfolio. Its distinctive technical claim is that the reasoning layer and execution layer are separate. An LLM can interpret filings or develop a thesis, while a deterministic event-driven strategy watches markets and handles execution without calling a model on every price change. The YC profile describes that architecture and the current supported venues.
The fit is a quantitative researcher, trading team or technically capable investor who wants agent workflows around market data without stitching together data vendors, compute, broker APIs and portfolio state. Scalar Field currently names Robinhood, Public.com, Polymarket, Jupiter DEX and Alpaca, and says paper trading is available through Alpaca and Polymarket. That makes the first evaluation concrete: start in paper mode, inspect the data and risk path, then decide whether live capital is appropriate.
Why I’d look closer
The rare useful detail is the boundary between “agent” and “trade.” Scalar Field’s launch description says the platform handles financial data, Python strategy development, backtests, persistent agent state, risk and execution checks, order routing, reconciliation and performance/NAV tracking. It also reports approximately 300ms event-to-trade latency as a company claim. The promise is not simply a chatbot that suggests a trade; it is infrastructure for repeatable strategy behavior.
The founders have relevant market and systems backgrounds. The official biographies describe Amandeep Singh as a former Tower Research and Goldman Sachs trader, and Ramakant Yadav as a former engineering leader at Microsoft. The current homepage fetch did not expose additional readable product detail in the bounded source set, so the YC profile remains the substantive public description checked here.
What I’d ask
What broker permissions are required, how are capital limits and kill switches enforced, and what is logged when a strategy diverges from its backtest? I’d inspect paper-trading parity, market-data provenance, slippage assumptions, reconciliation and audit history. The sources checked did not expose public pricing or a compliance boundary; those are not optional questions for a system that can execute trades.
My editorial take
Scalar Field is interesting as agent infrastructure for markets, not as a promise of automated returns. The architecture earns a closer look because it puts deterministic controls around the model’s less predictable reasoning. I’d judge it by reproducibility, failure handling and risk controls before the latency headline.
Quick facts
| Field | Sourced detail |
|---|---|
| Buyer fit | Quantitative researchers, trading teams and technically capable investors |
| Supported venues | Robinhood, Public.com, Polymarket, Jupiter DEX and Alpaca; company-stated |
| Safer first test | Paper trading through Alpaca and Polymarket, per company profile |
| Public pricing | Not exposed in the sources checked |
Sources checked
Checked 2026-09-19.
| Source | Used for |
|---|---|
| YC company profile | Product architecture, venues, founders and company-reported latency |
| Scalar Field homepage | Current public fetch; no additional readable product detail retained |
Cohort context
Scalar Field is listed in Spring 2025. In our 2026-09-18 directory snapshot, 7 of 143 listed companies in that cohort have YC’s primary industry label Fintech (4.9%). This is a current-directory comparison, not an original intake count or a performance ranking. Nine-cohort dataset.
Public website snapshot
Observed 2026-09-19T16:15:51.602Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.
| Signal | Homepage observation |
|---|---|
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Not observed in this response |
| Typed structured data | Observed |
| Docs/developer link | Not observed in this response |
| Pricing link | Not observed in this response |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |
Public observations · Collection method. Missing links here do not establish that a capability or file is absent elsewhere.
About the author
Cofounder of Lazyweb (a16z Speedrun). Previously grew subscriptions and built native ads at Duolingo. Writes First1000 (90,000+ subscribers).
I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.
First1000 ↗ · X ↗