About the Role
Build, train, deploy, and operate the ML engine powering TourIQ, a B2B dynamic pricing platform for the tours & activities industry. You'll work across traditional ML, LLM integration, and full MLOps no separate MLOps hire, you own the complete lifecycle from development through production. Your models directly impact customer revenue.
Key Responsibilities
ML Pricing Engine
- Design, train, and deploy ML models for price optimization using diverse real-world signals
- Build ensemble approaches combining multiple model outputs with confidence scoring
- Engineer and maintain a large feature set spanning temporal, environmental, demand, and competitive signals
- Handle scenarios with limited historical data for new customers
- Implement guardrails and fallback logic for low-confidence predictions
- Train and manage ML models for diverse customer segments at scale
MLOps & Production Operations
- Experiment tracking and model versioning; production model serving
- Build automated retraining pipelines with scheduled and performance-triggered cycles
- Model monitoring, drift detection, and alerting in production
- Model rollback and recovery procedures
- Explainability: surface key factors behind each recommendation so users understand and trust outputs
- Ensure consistency between training and serving data
AI Assistant & LLM Integration
- Build an AI assistant using LLM APIs for natural language interaction with platform data
- Implement RAG for contextual retrieval and knowledge grounding
- AI agent workflows: query data, generate reports, explain decisions in plain language
Collaboration
- Partner with Data Engineer, Tech Lead, and UI/UX Designer across the product
Must-Have
- 3+ years applied ML/Data Science with production models (not just notebooks/Kaggle)
- Python with PyTorch and XGBoost/scikit-learn
- End-to-end ML pipelines: data prep, feature engineering, training, evaluation, deployment, monitoring
- MLOps: automated retraining, model monitoring/drift detection, versioning and rollback in production
- Regression models, ensemble methods, gradient boosting, time-series forecasting
- Production model serving experience
- Experiment tracking (MLflow, W&B, or similar); SQL/PostgreSQL for feature engineering
- LLM API integration (OpenAI, Anthropic/Claude, or similar) in production
Good to Have
- Dynamic pricing / revenue management / demand forecasting; LSTM/RNN; Reinforcement Learning; RAG; vector databases; A/B testing; CI/CD for ML models; PySpark
Mindset
Ships production models AND keeps them running. Full lifecycle ownership. Starts simple, scales with data. Revenue-driven. Explains ML to non-technical users. Startup mentality.
Why Join
Own the entire ML/AI stack from day one. Build models that drive measurable revenue lift. Traditional ML + LLM + MLOps. Onsite in Jaipur.
