Research Assistant - Compiler AI

New York University; Al Ruways Industrial City; On-site; AED 133,920 - 200,880; Full time

Benefits offered by this job

Housing subsidy

Job summary

New York University Abu Dhabi seeks a Research Assistant - Compiler AI to advance automatic code optimization, deep learning, compiler technology, and high‑performance computing. The role involves collaborating with faculty, researchers, and industry partners on real-world optimization challenges.

You will implement compiler passes in the Tiramisu framework, develop ML models to predict optimizations, and benchmark across multicore CPUs, GPUs, and distributed environments.

Qualifications

• Bachelor's degree in Computer Science or equivalent.
• Master's degree in CS or related discipline preferred.
• Internship, academic project, or professional experience with compilers or AI/ML.
• Strong programming foundation in complex software systems.
• Familiarity with compiler concepts and ML model development.

Responsibilities

• Support research at the intersection of compiler tech, AI, DL, and HPC.
• Extend the automatic code optimization module in the Tiramisu framework.
• Develop ML models predicting compiler optimizations for programs and hardware.
• Design representations for loops, computations, memory access, dependencies, and schedules.
• Generate training data by compiling and benchmarking across configurations.
• Evaluate optimization strategies like vectorization, parallelization, loop fusion, blocking, tiling.
• Measure execution time, speedup, memory behavior, scalability, hardware utilization.
• Compare generated schedules with manual optimizations and baselines.
• Investigate cases where transformations improve or reduce performance.
• Develop compiler passes to apply optimization sequences to unoptimized code.
• Test generated programs on multicore CPUs, GPUs, FPGAs, and distributed environments.
• Build reproducible experiments, benchmarking, and model-evaluation pipelines.
• Analyze results and contribute to technical reports and conference submissions.
• Collaborate with faculty, researchers, students, labs, and industry partners.
• Participate in meetings, discussions, demonstrations, and project planning.

Skills

Compiler development

Compiler optimization

Automatic code optimization

Machine learning

Deep learning

Artificial intelligence

Tiramisu compiler

High‑performance computing

Parallel computing

Multicore programming

GPU programming

FPGA computing

Distributed computing

Vectorization

Loop parallelization

Loop fusion

Loop blocking

Loop tiling

Data locality

Intermediate representations

Code generation

Performance benchmarking

Runtime prediction

C

Python

LLVM

TensorFlow

PyTorch

Research programming

Reproducible experimentation

Education

Bachelor's degree in Computer Science

Master's degree in Computer Science or related discipline

Tools

LLVM

Job description

Role Overview

Research Assistant - Compiler AI at New York University Abu Dhabi, Abu Dhabi, United Arab Emirates — an opportunity in Higher Education research focused on automatic code optimization, deep learning, compiler technology, and high‑performance computing.

Salary range: 12,000–18,000 AED per month (final offer confirmed with the employer).

Job type: Full‑time; Function: Research; Experience: Internship or relevant experience in compilers and artificial intelligence.

Key Responsibilities

• Support research at the intersection of compiler technology, artificial intelligence, deep learning, and high‑performance computing.
• Extend the automatic code optimization module used within the Tiramisu compiler framework.
• Develop machine learning models that predict effective compiler optimizations for different programs and hardware targets.
• Design features and representations for loops, computations, memory access patterns, dependencies, and execution schedules.
• Generate and organize training data by compiling and benchmarking programs under multiple optimization configurations.
• Evaluate optimization strategies such as vectorization, parallelization, loop fusion, blocking, tiling, and data‑locality improvement.
• Measure execution time, speedup, memory behavior, scalability, and hardware utilization.
• Compare automatically generated schedules with manually optimized implementations and compiler baselines.
• Investigate cases where individual transformations improve or reduce program performance.
• Develop compiler passes that select and apply optimization sequences to previously unoptimized code.
• Test generated programs on multicore CPUs, GPUs, FPGAs, and distributed computing environments where applicable.
• Implement reliable research software and maintain clear source‑code documentation.
• Debug compiler transformations, code‑generation issues, model predictions, and performance regressions.
• Build reproducible experimentation, benchmarking, and model‑evaluation pipelines.
• Analyze research results and contribute to technical reports, academic papers, and conference submissions.
• Collaborate with faculty members, research scientists, students, external laboratories, and industry partners.
• Participate in laboratory meetings, research discussions, demonstrations, and project planning.

Ideal Profile

• Bachelor's degree or equivalent in Computer Science.
• Master's degree in Computer Science or a related discipline is preferred.
• Internship, academic project, or professional experience involving compilers, artificial intelligence, or machine learning.
• Strong programming foundation with the ability to understand and modify complex software systems.
• Familiarity with compiler concepts such as intermediate representations, code generation, loop transformations, scheduling, and optimization passes.
• Understanding of machine learning or deep learning model development and evaluation.
• Knowledge of computer architecture, parallel programming, or high‑performance computing is advantageous.
• Experience with C, C++, Python, LLVM, TensorFlow, PyTorch, or comparable technologies is beneficial.
• Ability to perform systematic benchmarking and interpret performance measurements.
• Strong analytical skills with careful attention to implementation details, experimental validity, and reproducibility.
• Good written and verbal communication skills for research collaboration and technical documentation.
• Self‑motivated approach with the capacity to work independently and contribute effectively within a research team.
• UAE Nationals are encouraged to apply.

Skills Set

• Compiler development
• Compiler optimization
• Automatic code optimization
• Machine learning
• Deep learning
• Artificial intelligence
• Tiramisu compiler
• High‑performance computing
• Parallel computing
• Multicore programming
• GPU programming
• FPGA computing
• Distributed computing
• Vectorization
• Loop parallelization
• Loop fusion
• Loop blocking
• Loop tiling
• Data locality
• Intermediate representations
• Code generation
• Performance benchmarking
• Runtime prediction
• C
• Python
• LLVM
• TensorFlow
• PyTorch
• Research programming
• Reproducible experimentation

Why Join Us

This position offers practical experience in a specialized research area connecting modern compiler design with artificial intelligence. The successful candidate will contribute to technology intended for real users, partner companies, and research laboratories while gaining exposure to advanced optimization methods and heterogeneous computing architectures. The role provides opportunities to strengthen expertise in machine learning, compiler engineering, performance analysis, and academic publication.

Employment Terms

Competitive salary, housing support, educational subsidies for eligible children, access to advanced computing resources, and collaboration within an internationally connected computer science research environment.

About the Company

New York University Abu Dhabi is a global research university supporting advanced work across computer science, engineering, science, social sciences, humanities, and the arts. Its Modern Compilers Lab develops programming systems and optimization technologies that make high‑level software efficient across multicore processors, GPUs, FPGAs, and distributed hardware platforms.