NLP Engineer aa
0 views2+ years
Posted 4 days ago
- Devops
- Docker
- FastAPI
- Git
- Llm
- Python
About the Role
This position is listed on behalf of a partner company, which manages all applications and next steps. Our partner is looking for an NLP Engineer based in India.
As an NLP Engineer, you will develop, train, and optimize advanced language models powering intelligent conversational experiences and AI-driven products. You will work across the entire machine learning lifecycle, from dataset preparation and model fine-tuning to evaluation, deployment, and production monitoring. Your focus will be on adapting language models to evolving conversational scenarios, improving response quality, and ensuring reliable and efficient model performance. You will experiment with different models, prompting strategies, and inference technologies to balance accuracy, latency, and operational costs. Working closely with engineering and product teams, you will transform NLP research and experimentation into scalable, production-ready solutions.
Responsibilities
- Model Development and Fine-Tuning: Train, adapt, and fine-tune language models (e.g., LLaMA, Mistral, Gemma, Qwen) using parameter-efficient fine-tuning (PEFT/LoRA) and full fine-tuning techniques to support conversational tasks.
- Evaluation and Benchmarking: Establish automated and qualitative evaluation workflows to measure model performance, safety, coherence, and alignment with product guidelines.
- Prompt Engineering and Structured Output: Design and iterate on prompt structures, tool-use interfaces, and JSON/structured output generation for production workflows.
- Production Deployment and Optimization: Deploy models into production environments using inference frameworks such as vLLM, TensorRT-LLM, or Triton, optimizing for low latency and high throughput.
- API Development: Build and maintain performant backend APIs (e.g., using FastAPI) to expose NLP and LLM services to client applications.
- Infrastructure and Collaboration: Collaborate with platform and DevOps engineers to maintain GPU-based inference infrastructure, monitoring systems, and deployment pipelines.
- Continuous Improvement: Monitor model behavior in production, conduct error analysis, and iterate on models to resolve failure modes and improve conversational quality.
- Data Strategy and Curation: Guide dataset curation, synthetic data generation, and filtering to improve data quality for training and evaluation.
- Recommendation Systems: Contribute to recommendation-related capabilities and explore opportunities to apply NLP and machine learning techniques to improve product experiences.
Requirements
- Strong proficiency in Python, with experience writing clean, maintainable, production-quality code beyond research notebooks.
- Practical experience with PyTorch and Hugging Face Transformers, including a solid understanding of transformer architectures, tokenization, model training, and text generation.
- Hands-on experience adapting and fine-tuning language models through supervised fine-tuning (SFT) and parameter-efficient approaches such as LoRA.
- Experience evaluating generated text using benchmarks, automated evaluation methods, and human review.
- Practical experience working with model APIs, prompt engineering, and structured output generation.
- Experience developing and maintaining APIs using FastAPI or an equivalent backend framework.
- Familiarity with Docker, Linux, and Git, including experience deploying machine learning models on GPU infrastructure.
- Experience implementing logging, monitoring, automated tests, and production troubleshooting procedures.
- Practical experience with recommendation systems, even at an introductory or project level.
Benefits
- Fully remote working arrangement for candidates based in India.
- Generous Annual Leave: 28 calendar days of paid vacation per year.
- Children's Education Support: Reimbursement of 50% of kindergarten or school expenses, up to $1,000 gross per year per child.
- Healthcare Support: Health insurance benefits or reimbursement of up to $1,000 gross per year per employee.
- Remote Workplace Support: Equipment reimbursement of up to $1,000 gross once every three years.