AI/NLP Engineer

Role Overview:

We are looking for a technically strong AI/NLP Engineer to design and build scalable conversational AI systems using a layered, hybrid approach. You will develop classical ML and rule-based NLP pipelines for intent recognition, entity extraction, and dialogue management, and escalate to fine-tuning and integrating large language models (LLMs) like LLaMA when higher-level understanding or generation is required.

Your work will enable efficient, maintainable, and cost-effective chatbot and AI applications by combining the best of traditional NLP/ML and state-of-the-art LLM capabilities.

Key Responsibilities:

• Design, develop, and optimize rule-based NLP and classical ML pipelines (intent classification, slot filling, pattern matching) for conversational AI.

• Fine-tune and adapt large language models (e.g., LLaMA) using LoRA, PEFT, and instruction tuning for complex or ambiguous use cases.

• Build and maintain retrieval-augmented generation (RAG) pipelines using vector search engines (FAISS, Pinecone).

• Architect modular chatbot systems combining classical ML, rule-based, and LLM components for robust and scalable solutions.

• Preprocess and curate datasets for both classical ML training and LLM fine-tuning.

• Optimize GPU-based training workflows with tools like Hugging Face Accelerate, DeepSpeed, and manage distributed training.

• Develop and execute evaluation metrics and testing protocols to assess both classical models and LLM outputs.

• Collaborate with engineering teams to deploy conversational AI models into production systems.

• Stay current with latest research and best practices in NLP, ML, and LLM fine-tuning.

Required Skills & Experience:

• 2 years experience with NLP, classical ML, and LLM fine-tuning.

• Strong proficiency in classical ML algorithms (SVM, logistic regression, decision trees) and NLP techniques (regex, rule-based parsing, intent/entity extraction).

• Hands-on experience fine-tuning large language models (LLaMA, GPT, Falcon) using LoRA, PEFT, or related methods.

• Proficient with NLP and ML libraries: scikit-learn, spaCy, Hugging Face Transformers, NLTK.

• Experience building retrieval-augmented generation pipelines with vector databases (FAISS, Pinecone).

• Familiarity with distributed GPU training and memory optimization techniques.

• Strong Python programming and software engineering skills.

• Ability to design modular, maintainable ML/NLP systems combining different techniques.

• Experience working in cross-functional teams and clearly communicating technical concepts