Member of Technical Staff, Machine Learning & Optimization

Pepr AI•San Francisco, CA

11d

About The Position

Pepr AI is building the AI Operator for growth to replace traditional ad agencies. They apply the quant rigor of high-frequency trading to optimize ad spend, delivering 20-50% upside in spend efficiency. They are managing spend for global category leaders like Cider and Cupshe with a clear path to managing billions of dollars. They are backed by Quiet Capital and are looking for early engineers. The Role is for a Machine Learning Engineer to build the models, optimization systems and algorithms that drive their autonomous decision engine. This role involves not just building models, but designing the core financial strategies that dictate how millions of dollars are deployed and the execution layer that carries them out. The goal is to find inefficiencies in the ad markets, translate them into automated trading strategies and build the write-path integrations that execute those decisions in the real world. They are looking for builders who can bridge the gap between theoretical research and production systems, who are pragmatic enough to ship an 80% solution in a week to capture immediate value but disciplined enough to evolve that solution into a robust, generalized platform that manages risk at scale.

Requirements

4+ years of experience applying machine learning and optimization to real-world problems.
Comfortable with classical ML, regression analysis and control theory.
Extensive experience deploying models into high-throughput production environments.
Knowledge of how to handle API rate limits, retries and failure states.
Strong grasp of probability, statistics and linear algebra.
Ability to read a research paper and implement it.
Prioritization of velocity and P&L impact.
Ability to ship fast to validate hypotheses and iterate based on live feedback from the market.

Nice To Haves

Rigorous course in Convex Optimization and understanding of how to formulate and solve complex constraint problems.
Experience as an accelerator using tools like Cursor, Claude Code or Codex to write boilerplate and tests.
Experience in quantitative finance, algorithmic trading or programmatic advertising (RTB).

Responsibilities

Design the Core Algorithms: Develop value models that predict customer LTV, risk-aware optimizers that allocate budget across channels and detection systems that identify creative fatigue before it kills performance.
Build the Execution Layer: Write integrations that execute optimal bids, including API connectors for Meta, Google and TikTok to robustly update bids, budgets and targets in real-time.
Deploy and Monitor at Scale: Own the model lifecycle from notebook to production. Build the observability infrastructure to detect concept drift, monitor inference latency and ensure trading decisions remain stable as market conditions change.
Backtest and Verify: Build simulation infrastructure to prove strategies work before capital is deployed. Ensure logic is robust to market volatility and generalizes across different client verticals.