Data Scientist - Artificial Intelligence/Machine Learning

The Role :

- You will do real data science on data that matters.

- Not Kaggle datasets.

- Not classroom assignments.

- Live production data millions of search queries, bookings, cancellations, reviews, and pricing signals from one of India's most active travel platforms.

- The models you build will serve real users.

- The analyses you produce will inform real product decisions.

- Work with live production data across all six verticals every dataset is connected to real outcomes

- Build models that progress from notebook to staging deployment with team support and code review

- Learn production ML practices: versioning, evaluation, monitoring, and iteration in a real engineering org

- Contribute to NLP and LLM experiments on real user query and review data

- This role is designed to make you a Staff Data Scientist in 23 years

Core Responsibilities :

Exploratory Analysis & Insight :

- Build analytical dashboards and reports that product managers and business leaders actually use to make decisions

- Identify data quality issues in production datasets and collaborate with Data Engineering to resolve them

Model Development :

- Build and evaluate ML models : classification (fraud, intent, complaint routing), regression (demand, price), clustering (user segmentation), and ranking baselines

- Implement basic feature engineering pipelines and contribute features to the central feature store

Train NLP models: text classification, intent detection, entity extraction, basic embedding models

- Support LLM pipeline experiments: build evaluation test sets, measure retrieval quality for RAG pipelines, document hallucination patterns

- Contribute to recommendation model evaluation: run offline metrics, produce holdout analysis, compare baselines

Experimentation Support :

- Assist in A/B experiment setup: metric selection, sample size calculation, monitoring dashboards

- Analyse experiment results: SRM checks, statistical significance, confidence intervals, business interpretation

- Build analysis notebooks that are reproducible and documented your exploratory work is a handover, not a throw-away

What You'll Learn On The Job :

Production ML :

- How models move from notebook staging production with tests, canary releases, and monitoring

Experiment Design :

- How to design statistically valid A/B tests in a live marketplace with network effects and selection bias

Feature Stores :

- How Feast works, how real-time features are served, and how feature drift is detected

LLM Systems :

- How RAG pipelines are evaluated, how prompt quality is measured, and how agentic systems are tested

Voice AI :

- How ASR/TTS systems work for Indian languages, and how spoken intent is classified at scale

Distributed Data :

- How Spark and Kafka power ML feature pipelines at hundreds of millions of events per day

Who You Are :

- 3 to 5 years in Data Science, ML, or a quantitative analytical role or a strong advanced degree with relevant project experience

- Proficient in Python (pandas, NumPy, Scikit-learn) and SQL your code is readable and documented

- Working NLP knowledge : text preprocessing, TF-IDF, embeddings, or basic fine-tuning LLM curiosity is highly valued

- Intellectually honest : you report null results cleanly and you do not overfit a story to noisy data

- Degree in CS, Statistics, Mathematics, Engineering, or equivalent quantitative field