Dean Stratakos

YC-backed

• Built in-house coding agent leveraging AWS Bedrock primitives that was used by devs, traders, and quants across the firm. • Designed, developed, and optimized high-performance trading system components in C++ and Python, including order-placing robots, asset pricing engines, market data bridges, and message transport infrastructure. • Worked with traders and quantitative researchers to build road maps for larger projects. Coordinated with the trading operations team to develop release and rollback plans, ensuring safe deployment of production code. • Ran 2024 summer internship, managing 24 dev interns and 11 other mentors. Presented on system architecture and C++, scoped and assigned tasks, and reviewed code. Also mentored two interns during 2024 winter internship.

CS 142 - Web Applications • Led discussion sections and assisted students on projects during office hours. • Built autograder and deployed to Gradescope for efficient grading of projects.

• Coordinated three new Siri personalization features across five teams.

• Met one-on-one with a high schooler to prepare for USACO Silver (http://usaco.org). • Taught Python, C++, and JavaScript fundamentals to a middle schooler.

• Redesigned the recovery mechanism between market gateway and market connector nodes on Citadel’s internal trading platform. • Collaborated with traders to respond to several urgent production issues.

• Coded a matching algorithm for Athlete Mingle, a social event to bring together the athlete community. https://github.com/dastratakos/AthleteMingle • Contributed ideas to committee meetings with 15 other members. • Communicated between the student-athlete community and the SAAC organization, primarily with the Stanford Men's Tennis team • Provided behind-the-scenes support during annual events. SAAC Social Events hosts two annual events with the goal of fostering positive interactions among Stanford student-athletes. The first event is Stanford's Got Talent, a comedic and memorable talent show where student-athletes perform with their teammates. The second event is Stanny's, an awards banquet showcasing the most outstanding achievements of the year within Stanford Athletics.

• Computed per-pixel parallax values for videos shot on iPhone using LiDAR depth data. Developed visual representations using Matplotlib and OpenCV. • Implemented a homography estimation algorithm to help identify outliers within the "Effect Suggestions" system.

• Improved the performance, scalability, and maintainability of a machine learning clustering algorithm that grouped duplicate kernel panic reports. • Achieved cluster efficiency ARIs of 84-89% for 2 new panic signatures on iOS, macOS Apple Silicon, and macOS Intel. • Concepts included agglomerative clustering, tf-idf, and cloud storage.

• Implemented the backend for six CNN layers in a C++ based intermediate language designed for optimal performance on a specialized edge-computing hardware architecture at this startup. • Scrutinized compile-time optimizations and run-time optimizations for a deep learning network. • Studied post-training quantization of neural net models to improve performance during inference.