Forward Deployed AI Engineer
BCG X
- Building GenAI solutions on retail media data, working directly with client teams to ship deployed systems.
Resume · 2026
Ethan Lin
AI Engineer @ BCG X | CS @ UChicago
Interests GenAI Deployment · AI Infrastructure · Parallel Computing · GPU Programming · Algorithms
01 · Education
Education
ExperienceSep 2025 – Dec 2026
University of ChicagoGPA 3.92 / 4.0
Master of Science, Computer Science
Algorithms · Functional Programming · Distributed Systems · Parallel Programming · GPU Programming · Advanced C++
Chicago, IL
Jan 2021 – Dec 2024
University of New South WalesGPA 82 / 100
Bachelor of Science, Computer Science
Bachelor of Commerce, Business Analytics
Computer Systems · Data Structures · Algorithms · Operating Systems · Analytics · Statistical Modeling
Sydney, Australia
02 · Experience
Experience
GitHubCourse Assistant
University of Chicago, Department of Computer Science
- Designed programming assignments for a course of 100+ students, covering algorithms topics including randomization, dynamic programming, graph algorithms, min-cost max-flow, Bellman-Ford SPFA, and vertex cover.
- Built Python test suites and GitHub CI/CD workflows that automated grading and validation, removing manual checks from the weekly assignment cycle.
Library Student Assistant
University of Chicago
- Built an LLM-assisted pipeline using structured-output prompting to convert images into standardized catalog metadata, replacing manual data entry in the cataloging workflow.
- Developed Playwright automation tools that eliminated repetitive manual steps from cataloging workflows.
Software Engineering Intern
Circana
- Optimized SQL logic for consumer-insight reporting, reducing analytics data generation time by 40%.
- Built a black-box test suite for query validation, reducing manual validation cycles by 20%.
- Implemented C# features for the Liquid Data platform used by 20+ major retail partners.
- Maintained and refactored a 5,000+ line legacy codebase to improve reporting accuracy.
Data Analyst Intern
Tycoon Capital
- Built Scikit-learn pipelines on 100K+ ASX records, improving equity-price prediction accuracy by 25%.
- Converted model outputs into trading-strategy signals used by 3 portfolio managers.
03 · Projects
Projects
View GitHubCUDA Image Search Engine
CUDA · CLIP · Embeddings · GPU Programming
- Built a CUDA-accelerated image similarity search engine using CLIP embeddings and custom GPU kernels for cosine similarity search over ~7,400 images.
- Designed a multi-stream CUDA pipeline to overlap memory transfers, kernel execution, and result write-back across batches.
- Profiled and tuned kernel/batch parameters, achieving ~200× speedup over a CPU baseline with reproducible benchmarking scripts.
Matrix Engine
C++20 · Template · Parallel Computing
- Applied C++ template programming to build a type-safe matrix computation library with compile-time dimensions and move-only ownership semantics.
- Designed sequential and parallel execution modes using lock-based/lock-free executors, futures, atomics, and task queues.
- Implemented and benchmarked parallel add, multiply, and determinant operations, achieving ~3× speedup over a sequential baseline.
Parallel K-means Clustering
Go · BSP · MapReduce · Slurm
- Implemented K-means in Go with sequential/BSP/MapReduce execution models.
- Optimized parallel point assignment, per-worker local reductions, and lock-free work stealing, reducing end-to-end runtime by 40%.
- Automated Slurm-based HPC benchmarking with Bash scripts across thread counts, dataset sizes, and execution modes.
FaaS Platform
FastAPI · Redis · ZeroMQ · Python multiprocessing
- Built a Function-as-a-Service platform supporting 500+ concurrent requests.
- Scaled execution throughput to ~1,000 functions/sec using multiprocessing workers.
- Integrated Redis for low-latency state, caching, and coordination to improve data access and fault tolerance.
SVF Static Analysis
C++ · LLVM IR · Software Security
- Developed a C++ static-analysis detector for the open-source SVF framework.
- Used interprocedural control-flow analysis and abstract execution to identify vulnerability patterns.
- Wrote 100+ C and LLVM IR test cases to benchmark correctness and performance.
Epidemic Simulation Server
PostgreSQL · Flask · SQLAlchemy · Agile
- Designed a PostgreSQL schema modeling relationships across 1M+ records.
- Built a Flask server running 10 concurrent epidemic simulations with 40% lower latency after SQLAlchemy query optimization.
- Managed roadmap planning for a cross-functional team of 5 under Agile development.
04 · Skills
Technical skills
Contacti · Languages
Languages
- Python
- C / C++
- CUDA
- Java
- C#
- SQL
- Haskell
- Bash
ii · Frameworks
Frameworks
- Flask
- React
- PostgreSQL
- Redis
- Docker
- AWS
- ZeroMQ
iii · Data & Tools
Data & Tools
- pandas
- NumPy
- Scikit-learn
- TensorFlow
- GPU Programming
- Nsight Compute / Systems
- Linux
- Git
- pytest
- Power BI
- Jira