TF

Tianjun (Robbie) Feng

MSc. Computer Engineering at Columbia University | AI & Machine Learning Engineer

About Me

I am a Machine Learning and Software Engineer pursuing my Master's in Computer Engineering at Columbia University. My expertise lies in building intelligent agents, optimizing high-concurrency backend systems, and leveraging Large Language Models (LLMs) for complex tasks. From automating Android agents at OPPO to designing financial intelligence systems, I specialize in bridging the gap between advanced AI research and scalable production applications.

Python Java C/C++ Spring Boot Docker AWS & GCP Machine Learning Unity

Experience

Machine Learning Engineer Intern OPPO Jul 2025 - Aug 2025 | Shanghai

Contributed to a mobile intelligent agent for executing complex NL commands on Android. Accelerated model iteration by 25% and benchmarked performance against SOTA frameworks like Mobile-Agent-v2.

Software Engineer Intern Jingdong Logistics May 2024 - Jul 2024 | Shanghai

Developed high-concurrency backend systems for the JD.com e-commerce platform. Built idempotent APIs and stress-testing tools to support high-availability coupon services and promotion workflows.

AI Engineer Intern China Fortune Securities May 2023 - Jul 2023 | Shanghai

Engineered an automated Financial Intelligence System using GPT-4 APIs and Prompt Engineering. Optimized information retrieval by developing custom web crawlers for real-time market news analysis.

Selected Projects & Research

GlitchBench (CVPR 2024)

Research on detecting video game glitches using Large Multimodal Models. Developed a Unity program to automate training sample generation, writing over 1500 lines of C# script to identify object positions and bugs.

View Publication →

Autonomous Android Agent

Designed an intelligent agent capable of cross-app information retrieval and multi-step planning. Evaluated on the Android Agent Arena (A3) to analyze UI detection accuracy and planning stability.

View Details →

Financial Intelligence System

An automated system mapping unstructured news data to economic indicators using GPT-4. Collaborated with quantitative analysts to align AI solutions with high-frequency trading frameworks.

View Details →

Let's Connect

I am currently based in New York and open to new opportunities. Whether you have a question or just want to say hi, my inbox is always open.

tf2637@columbia.edu