Leads the team you would join
Owns the opening
On the team, can refer you
From your own Gmail and LinkedIn.
About the role
This is a data scientist role on a growing analytics team, onsite in New Delhi, India, for someone with two to four years of hands-on data science or machine learning experience. It suits people who have deployed models in production and want to apply transformer, embedding and retrieval methods to healthcare and clinical trial data. The work spans the full machine learning lifecycle on Databricks, from data ingestion and experimentation through deployment, monitoring and retraining. Datasets include patient records, trial protocols and adverse event data, and the role works closely with domain experts, product managers and engineers.
What you’ll do
- Design, build and tune predictive models across the full ML lifecycle.
- Run exploratory data analysis and guide feature engineering choices.
- Validate model performance with suitable statistical methods and domain knowledge.
- Deploy and monitor production models on Databricks MLflow endpoints using Unity Catalog.
- Build CI/CD pipelines for model versioning, experiment tracking and automated retraining.
- Keep models reliable, observable and performing well in live environments.
- Apply BERT-family and ClinicalBERT models to classification, named entity recognition and information extraction.
- Build and maintain vector similarity search for semantic retrieval and recommendations.
- Fine-tune pre-trained models for clinical and healthcare uses.
- Explore LLM integration and prompt engineering for internal tools.
- Turn ambiguous business problems into structured frameworks with measurable outcomes.
What they’re looking for
- Two to four years of hands-on data science or machine learning work.
- Proven deployment of ML models in production, not only prototypes.
- Strong Python with pandas, NumPy, scikit-learn and PyTorch or TensorFlow.
- Experience with Databricks, MLflow (experiment tracking, model registry, endpoints) and Unity Catalog.
- Hands-on work with BERT-family models and Hugging Face Transformers.
- Familiarity with vector databases such as FAISS, Pinecone or Weaviate, and embedding-based retrieval.
- Solid SQL and experience with large structured and unstructured datasets.
- Exposure to AWS, GCP or Azure, and to distributed computing such as Spark.
- Prior work with clinical trial data standards such as CDISC, CDASH and SDTM, or healthcare ontologies such as SNOMED and ICD-10.
- Familiarity with Trial2Vec or similar trial-to-vector embedding methods.
- Experience with LLM fine-tuning, RAG pipelines or prompt engineering in production.
- Knowledge of healthcare AI regulation and compliance, such as FDA guidance and HIPAA.
What’s on offer
- Onsite, based in New Delhi, Delhi.
Skills
Questions about this role
Where is this job?
New Delhi, Delhi, India, on an onsite basis.
How much experience is needed?
Two to four years of hands-on data science or machine learning experience, including models deployed in production.
Is Databricks experience required?
Yes, the posting asks for experience with Databricks, MLflow and Unity Catalog.
Do I need healthcare experience?
Prior work with clinical trial data standards or healthcare ontologies is listed as a requirement, along with familiarity with healthcare AI regulation.