principal data scientist

Location: MakeMyTrip, Bangalore

Experience Level: 8-10 years

Position: Principal Data Scientist

Reimagine Travel decisions with the power of AI @ MakeMyTrip. MakeMyTrip has been AI/ML/Data/Platform powered organization. We are now shaping the future of travel for Bharat customers, with GenAI-centric innovations, leveraging SLMs, Vernacular STT/TTS, and travel content systems. Join us to take these systems, chatbots, Cx, content to next orbit of excellence.

About The Role

As a Principal data scientist , you will:

• Design, develop, and fine-tune multilingual Small Language Models (SLMs) with advanced AI capabilities for agentic applications such as travel destination expertise, Myra bot orchestration, hotel semantic search, and customer support bot assistance.
• Build custom Natural Language Inference (NLI) and reward-scoring models to improve preference-based SLM training (DPO, PPO, SteerLM etc), enhancing multi-agent response quality and relevance.
• Develop and implement active learning pipelines for data and models, for NLP, Named Entity Recognition (NER), conversational latent intent/concepts, travel-specific embeddings, and domain-specific SLM training for summarization, rephrasing, and planning .
• Architect and enhance conversational AI systems to dynamically adapt and comprehend concepts/intents, NLP, and NER tasks across multiple lines of business, supporting the Myra chatbot ecosystem.
• Drive end-to-end quality assurance through robust integration and experience testing, establishing objective evaluation metrics and leveraging A/B testing for performance optimization.
• Mentor and guide engineering teams in advanced SLM and Generative AI analytics , fostering growth and innovation.
• Collaborate across teams to deliver scalable solutions that impact millions of users, ensuring alignment with business goals.
• Focus on hosting and deploying Generative AI models on robust and efficient infrastructure to support large-scale operations.
  
  

• Deep experience in Natural Language Processing (NLP & NLI), tiny/small language models, base language model retraining (both transformer and BERT based architectures).
• Extensive experience in conversational NLP systems.
• Experience in SLM adaptor fine tuning, SFT training, SLM pruning (base and instruct models)
• Experience in retriever frameworks in Q&A bots.
• Experience in contrastive learning, graph learning based embedding methods with heterogenous features.
• Preferably experience in building NLP active learning pipelines.
• Demonstrable experience in scaling your models to large scale search systems is a plus.
• Very good critical reasoning skills, data centric NLP/NLI systems.
• Must have temperament and hunger to work in fast-paced environment.