Matrimony.com - Artificial Intelligence Engineer - NLP/LLM

Description :

Job Summary :

We are looking for a passionate AI Engineer with strong expertise in Python, Natural Language Processing (NLP), and Large Language Models (LLMs).

Youll work on developing intelligent language-based solutions, integrating LLM APIs (like OpenAI, Cohere, Claude, Hugging Face), and building custom NLP pipelines for real-world applications.

Key Responsibilities :

Model Development & NLP Engineering :

- Design and build NLP pipelines for tasks like classification, NER, summarization, sentiment analysis, and Q&A.

- Fine-tune and optimize open-source LLMs (e.g. , BERT, GPT, T5) using Hugging Face Transformers.

- Preprocess and structure large-scale textual datasets using SpaCy, NLTK, or custom tokenizers.

LLM Integration & Prompt Engineering :

- Integrate LLM APIs (OpenAI, Claude, Cohere, Hugging Face) into real-time applications.

- Develop and iterate on prompt strategies to maximize LLM performance and reliability.

- Collaborate with product teams to embed LLMs into chatbots, match explainers, recommendation systems, and user-facing tools.

Deployment & Continuous Improvement :

- Deploy models as APIs using Flask, FastAPI, or Streamlit; containerize with Docker.

- Monitor model outputs and performance metrics; iterate based on evaluations and user feedback.

- Stay updated with the latest in GenAI, NLP benchmarks, and open-source tools.

Required Skills & Experience :

- Proficient in Python and common NLP libraries (SpaCy, Transformers, NLTK, Gensim).

- Experience integrating LLM APIs and building prompt-driven applications.

- Hands-on with Hugging Face Transformers for training/fine-tuning.

- Understanding of key NLP concepts : embeddings, attention, tokenization.

- Familiarity with vector DBs like FAISS, Weaviate, or Pinecone (for RAG).

- Comfortable with Git, containerization (Docker), and basic REST API development.

Preferred / Bonus Skills :

- Experience with LangChain, LlamaIndex, PromptLayer, or similar LLMOps tools.

- Knowledge of RAG workflows, zero/few-shot techniques.

- Familiarity with evaluation metrics like BLEU, ROUGE, perplexity, cosine similarity.

- Exposure to AWS, GCP, or Azure for cloud deployment.

- Contributions to NLP open-source projects or research work is a plus.