Software Engineer - LLM

Description :

About the Role :

We are seeking a Software Engineer with expertise in Large Language Models (LLMs) to join our AI/ML team.

The ideal candidate will work on designing, developing, and deploying state-of-the-art LLMs for real-world applications.

You will collaborate with research scientists, data engineers, and product teams to build scalable AI solutions that leverage natural language understanding, generation, and reasoning.

Key Responsibilities :

- Develop, fine-tune, and deploy Large Language Models (LLMs) for applications such as chatbots, summarization, recommendation systems, and search.

- Work with transformer-based architectures (e.g., GPT, BERT, T5, LLaMA) to implement and optimize LLMs for various tasks.

- Design and implement data pipelines for preprocessing, cleaning, and augmenting large-scale textual datasets.

- Optimize model performance for latency, throughput, and memory efficiency for production deployment.

- Collaborate with ML researchers to implement novel algorithms and improve model accuracy, reasoning, and generalization.

- Integrate LLMs into applications via APIs, microservices, and cloud-based platforms.

- Participate in code reviews, testing, and deployment pipelines to ensure high-quality, maintainable, and scalable software.

- Stay updated with the latest advancements in LLMs, NLP, and AI research and apply relevant techniques to projects.

Technical Skills :

- Strong programming experience in Python, with proficiency in PyTorch, TensorFlow, or JAX.

- Hands-on experience with transformer architectures and LLM frameworks such as Hugging Face Transformers, OpenAI API, or DeepSpeed.

- Experience with fine-tuning, prompt engineering, and model evaluation techniques.

- Knowledge of tokenization, embeddings, attention mechanisms, and sequence modeling.

- Experience in deploying models using cloud platforms (AWS, GCP, Azure) and containerization tools (Docker, Kubernetes).

- Familiarity with data processing frameworks such as Pandas, NumPy, PySpark, or Dask.

- Experience in training large models on GPU/TPU infrastructure and optimizing model inference.

- Understanding of MLOps practices, including CI/CD pipelines for AI models