Hotelzify

Role Overview

As an AI/ML Engineer at Hotelzify, you will be at the forefront of designing and deploying AI components that power our conversational agent (voice + chat). You’ll be responsible for NLP/NLU pipelines, real-time inference optimization, dialogue management, retrieval-augmented generation (RAG), and LLM integration, helping the agent understand user intents, answer questions, and complete bookings live.

Key Responsibilities

• Design and implement NLP/NLU models to understand real-time user intents from text and voice.
• Build and fine-tune LLM-based conversational flows using RAG, prompt engineering, and retrieval mechanisms.
• Integrate external tools for hotel availability, pricing APIs, CRM data, and transactional workflows.
• Develop efficient real-time inference pipelines with latency under 300ms for voice and chat.
• Collaborate with frontend/backend teams to ensure seamless LLM API orchestration.
• Optimize prompt logic, dialogue memory, and fallback strategies for natural conversations.
• Conduct A/B experiments and continuous learning pipelines for feedback-driven improvement.
• Use vector databases (e.g., FAISS, Pinecone, Weaviate) for retrieval over hotel-related data.
• Work on voice-specific challenges: STT (Speech-to-Text), TTS, and intent detection over audio streams.

Tech Stack

• Languages: Python, Node.js
• AI/ML: LangChain, Transformers (Hugging Face), OpenAI APIs, LlamaIndex, RAG, Whisper, NVIDIA NeMo
• Infra: AWS (EC2, RDS, EKS, Lambda), Docker, Redis
• Databases: PostgreSQL, MongoDB, Pinecone / Weaviate / Qdrant
• Voice APIs: Plivo, Twilio, Google Speech, AssemblyAI

What We’re Looking For

• 3+ years in AI/ML/NLP-focused roles, preferably in production environments.
• Strong understanding of modern LLM pipelines, LangChain/RAG architectures.
• Experience building or integrating real-time conversational AI systems.
• Comfortable with voice-based systems: STT, TTS, and real-time latency tuning.
• Hands-on experience with fine-tuning or prompt-tuning transformer models.
• Bonus: Experience working in travel, hospitality, or e-commerce domains.

Nice to Have

• Prior work with agents that use [tool calling] / [function calling] paradigms.
• Knowledge of reinforcement learning for dialogue optimization (e.g., RLHF).
• Experience deploying on GPU-based infrastructure (e.g., AWS EC2 with NVIDIA).

Why Join Us?

• Work on a real product used by thousands of guests every day.
• Build India’s first real-time AI agent for hotel sales.
• Flexible work environment with deep ownership and autonomy.
• Get to experiment and deploy bleeding-edge ML/AI tech in production.