About Us
Secure Lasting Services builds bidding and trading software for renewable generators and storage operators in US wholesale power markets, starting with day-ahead bid optimization in CAISO.
Renewable bidding is, at its core, a statistics problem wearing a market's clothes. The optimal day-ahead commitment is a quantile of a production forecast distribution, chosen by the conditional distribution of the DA-RT spread and corrected for the covariance between the two. Getting that right demands calibrated probabilistic forecasting, honest backtesting, and optimization under uncertainty. That is precisely the work our founder has spent his career doing.
Our founder's path here was anything but linear: mechanical and electrical engineering as an undergraduate, a master's in computer science, more than a decade in software engineering, then a PhD in economics, specializing in Bayesian econometrics. When the current AI wave arrived, he was shipping production machine learning systems: first deep learning models, then two and a half years as a founding AI engineer building decision-support systems in a regulated industry. SLS began as an AI engineering consultancy built on that experience. In 2026 we focused the company on the market where rigorous probabilistic modeling pays its way most directly, every single day.
The economics training matters more here than in most industries, because US wholesale power markets were designed by economists. The price at every node is the shadow price of a welfare-maximizing optimization, an idea that traveled from peak-load pricing theory through spot-pricing research into the settlement software of every ISO. The capacity auctions, the scarcity pricing rules, the financial transmission rights: all of them began as economics papers before they became tariff provisions. We traced that history in a survey published in our resources: How Economists Have Shaped the US Electricity Markets. For an economist, this market is not an exotic domain to adapt to. It is the one industry built in the language he was trained in.
We came to power markets with a conviction formed by that background: the operator's core engine is optimization, not a black-box price oracle. Nobody reliably predicts prices. What you can do is model the distributions that matter, at the node where you settle, with information available before the gate closes, and make the bid that is optimal against those distributions. Then you measure the result honestly: out-of-sample, point-in-time, against a naive baseline, sliced by market regime.
That last part matters most in this industry. Vendor uplift claims are marketing figures on non-comparable baselines. Ours is a business built on the opposite premise: every number we show a client is one they can audit.