Lead Data Scientist - Machine Learning

About the role:

We are looking for a Lead Data Scientist who sets technical direction for a workstream, goes deep on the modeling and architecture decisions personally, and leads a small team of data scientists and engineers. You will be the senior technical authority the client turns to when a model decision needs defending.

What you will do:

- Own the full model lifecycle end to end: problem framing, feature engineering, model architecture selection, training at scale, validation, deployment, and post-launch monitoring and retraining.

- Make and defend architecture-level calls: which model family, how much complexity is actually justified by the data and the business case, when a classical model beats a deep learning one and vice versa.

- Design and run rigorous experiments (A/B tests, causal inference, uplift modeling) and be able to explain confounders and why an offline metric lied to you.

- Build and own feature pipelines and training infrastructure that hold up at production scale and under data drift.

- Diagnose model degradation in production and make the retrain-versus-redesign call, including rollback plans.

- Set the technical bar for the team: code review standards, experiment tracking, model validation rigor, and mentor 2-4 data scientists and engineers against it.

- Be the primary technical point of contact for the client, translating ambiguous problems into scoped work and defending model tradeoffs and failure modes to a non-technical audience.

Must-haves:

- 10 - 12 years of experience in data science, with demonstrated ownership of models from problem definition through production impact and measured business outcome.

- Deep, defensible grounding across model families: classical ML (regression, tree ensembles like XGBoost/LightGBM) and deep learning (architecture choice, training at scale), with clear judgment on when each is the right call.

- Domain depth in at least one deep learning area relevant to enterprise work: NLP, forecasting, recommendation systems, or computer vision, including hands-on architecture and training decisions, not just fine-tuning a pretrained model.

- Production-grade feature engineering and training infrastructure experience, including experience with distributed training or large-scale compute (Spark, Ray, or equivalent).

- Experience owning a model in production long-term: monitoring, drift detection, retraining triggers, rollback.

- Experience leading a team technically, including mentoring and setting the standard for others' modeling and code work.

- Strong client-facing communication, able to hold a technical argument with a client stakeholder and explain a model's limitations plainly.

- Cloud/ML platform stack - Databricks, SageMaker, Vertex AI, Azure ML, etc.

- GenAI/LLM applied experience: retrieval design, evaluation harnesses, honesty about failure modes, not just demo projects.

- MLOps tooling depth: MLflow or similar model registries, automated retraining pipelines, CI/CD for ML.

- Prior consulting or professional services background, comfortable across multiple concurrent client engagements.

Primary Skills:

- Classical ML - regression, classification, tree ensembles (XGBoost/LightGBM), model selection judgment.

- Snowflake - Should be genuine hands-on experience working on Snowflake Platform.

- Deep learning - architecture design and training at scale in at least one domain area (NLP, forecasting, recommendation systems, or computer vision).

- Causal inference / experimentation - A/B testing, uplift modeling, confounder-aware analysis.

- Feature engineering and training infrastructure - production-scale pipelines, distributed compute (Spark, Ray, or equivalent).

- MLOps / production ownership - deployment, drift and degradation monitoring, retraining triggers, rollback.

- Technical leadership - mentoring, code/model review standards, setting team technical bars.

- Client communication - defending model tradeoffs and limitations to nontechnical stakeholders.

- Python and SQL - production-grade.

Why Join KANINI?:

- People-first culture with diversity and inclusion at its core.

- Recognized as a Great Place to Work.

- Join KANINI's award-winning Data Engineering Team, recognized as the "Outstanding Data Engineering Team" at DES 2025.

- Exposure to cutting-edge technologies: AI, Data Analytics, Cloud, IoT, Telehealth.

- Opportunities for career growth, mentorship, and impactful projects in Healthcare and BFSI domains.

Ready to Make an Impact?:

- Contribute to impactful projects that shape the future of data and AI.

- Collaborate with top-tier professionals in a dynamic, fast-paced environment.

- Take ownership of your work and make a tangible difference in the company's success.

- Grow your career with mentorship, training, and opportunities for advancement.