Pluspunten
As a Senior Software Engineer, you eventually look for three things in a job: a complex technical problem to solve, a direct impact on the business, and a culture that trusts you to build. Upstart doesn't just check these boxes—it excels at them. Many companies call themselves "AI-first," but here, it’s not a marketing slogan; it’s our entire architecture. We're not just building a loan application website. We are engineering a sophisticated, multi-variable, real-time decisioning engine that is fundamentally replacing the archaic FICO score. From a technical standpoint, the challenge is immense. Our models have evolved from simple linear regressions to complex neural networks that analyze thousands of data points—far beyond what a traditional underwriter could ever handle. This isn't a "once-a-month" batch-processing job. This platform runs 24/7, with over 90% of our loans being fully automated. As an engineer, that means your code is responsible for billions of dollars in originations, and you have to build for massive scalability, high availability, and iron-clad security. The data pipelines alone, likely leveraging tools like Kafka, are a fascinating challenge in distributed systems. What’s most rewarding is the "why." Our work has a clear, measurable, and immediate impact. When the company announced its Q3 2025 results—reporting a sixfold increase in net income and 71% revenue growth—it was a direct validation of the platform we're building. Our AI model is our product, and it's winning. But the "why" is also about the mission. We're not just optimizing ad-clicks. We're building a system that demonstrably provides more equitable access to credit, approving far more applicants from minority backgrounds at lower interest rates than any traditional bank. The "digital-first" culture respects senior talent, giving us the autonomy to solve these problems without being bogged down by legacy bureaucracy. It's the rare place where you get to solve a Google-level AI problem with the tangible, real-world impact of a fintech leader.
Minpunten
The AI Model's "Overreaction" This is the single biggest "negative" right now. The core promise of Upstart is that its AI is smarter and more adaptable than traditional credit models. However, in its Q3 2025 earnings, management admitted its own AI model "overreacted" to macroeconomic signals. This caused the model to automatically "tighten the credit box" (becoming overly conservative), which led to: A Drop in Conversion: The loan conversion rate fell from 23.9% to 20.6%. Missed Opportunity: This happened at the same time that consumer loan applications were at a three-year high. The AI's "overreaction" made the company miss out on significant loan volume, which rattled investors. It calls into question the very stability and "all-weather" reliability that is central to Upstart's value proposition.