Arabic LLM Research Engineer

You will be part of a large team of research engineers and work on developing large language models specialized to meet the customers' specific needs and culture. You are expected to lead portions of this team in research and development tasks such as fine tuning (SFT = Supervised fine tuning), reinforcement learning (RLHF, DPO), retrieval augmented generation as well as more advanced tasks at the cutting edge of research. You are expected to provide feedback and direction to the data engineering team with respect to data collection, cleaning and processing, and also to collaborate with the annotation team with respect to the evaluation of the large language model. Being a senior engineer you will spend time leading and mentoring junior research engineers in the completion of their tasks. In addition to product based efforts you are expected to keep up with the scientific literature and contribute to it as well by publishing papers.

1. Basic Computer Science and Programming

· Data Structures and Algorithms (practice with LeetCode type problems).

· Object Oriented Programming.

· (optional) Functional Programming.

· Writing robust and reusable code according to design patterns.

· Ability and willingness to write readable, and well documented Python code.

· Version Control (git)

2. Machine Learning and Deep Learning

· Classical Machine Learning and Deep Learning especially as applied to NLP.

· Scikit-Learn, Jupyter Notebook/Lab, Pandas, Numpy

· At least one of Pytorch, Tensorflow (not Keras) or JAX.

· Ability to design and implement scalable neural models.

· Model Compression (Quantization, Distillation, Pruning)

· (optional) Reinforcement Learning

· (optional) Active Learning

· (optional) Weakly-supervised Learning

3. Natural Language Processing

· Data Collection and Preprocessing

· Text Representation (e.g. BoW, word embeddings, contextual embeddings)

· Neural Network Architectures used in NLP.

· Language Models (n-gram, masked and causal)

· Experience with common NLP tasks such as spelling correction, text classification, token classification (NER, PoS etc..), machine translation.

· (optional) Information Retrieval

4. Large Language Models

· Huggingface transformers

· In-context Learning (prompt engineering)

· Supervised Fine Tuning (SFT)

· Alignment (e.g. RLHF-PPO, DPO)

· 3D Parallelism (data, tensor, pipeline)

· Model Deployment (inference optimizations)

5. Academic Credentials

· Masters or Ph.D. in Computer Engineering or related fields.

· Ability to read and implement SoTA LLM papers.

· Willingness to follow NLP and LLM literature and stay informed of latest developments.

· Publications in reputable conferences and journals.

6. Non-technical Requirements

· Fluency in written and spoken English.

· (optional) Fluency inArabic, Portuguese, Spanish or Russian.

· Strong communication and leadership skills.

• A real job from day one: We offer you a professional career in one of the leading multinational technology company.
• Local & international: Reaching more than 190 countries, Huawei is a successful and respected business. We focus on the needs of local costumers by harnessing global expertise and team work. For you, that means exceptional exposure and experience.
• Great Development Opportunities: We'll support you every step of the way, with hands on experience which includes functional, cross functional and international rotations.
• Fast growth and ambitious vision.
• Learning and Development opportunities.