Senior Data Scientist

SvaaN Global Tech PVT LTD; Chennai; On-site; INR 2,000,000 - 3,200,000; Full time

Job summary

SvaaN Global Tech PVT LTD is seeking a Senior Data Scientist in Chennai to develop predictive and forecasting models for large business datasets. The role emphasizes production-ready solutions, scalable pipelines, and collaboration with data engineers and platform teams.

You'll work with neural networks, transformer-based architectures, and traditional ML to deliver models that influence decision-making and business optimization. Strong Python and ML tooling experience is required.

Qualifications

• 6+ years of experience in data science, applied ML, AI research engineering or related roles working with large datasets.
• Strong foundation in ML, statistics, probability, optimization, model evaluation and experimental design.
• Hands-on experience training and evaluating neural networks, deep learning models, sequence models or transformer-based models.

Responsibilities

• Develop predictive, forecasting and anomaly detection models for large datasets.
• Work with neural networks, transformer-based models, and traditional ML where appropriate.
• Define training labels, evaluation methods, and model performance criteria tied to business outcomes.
• Collaborate with data engineers to create reliable training datasets and reproducible model pipelines.

Skills

Python programming

ML foundations

Communication

Productionization

Education

Bachelor's/Master's in CS/Math/Stats

Tools

PyTorch

TensorFlow

scikit-learn

XGBoost

MLflow

Hugging Face

Job description

Role Summary

The Senior Data Scientist will develop models and analytical methods for business optimization using large datasets, neural networks, model training and high-scale transformer-based approaches where appropriate. This is a generic AI platform role and is not limited to a single industry or data domain.

The role requires someone who can move beyond notebooks and contribute to production model development. The right person has strong statistical and ML foundations, understands data quality and evaluation, and can create models that support predictions, forecasts, recommendations, risk scoring and decision-making at scale.

What You Will Do

• Develop predictive models, forecasting models, anomaly detection methods, classification/ranking models and optimization approaches for large business datasets.
• Work with neural networks, deep learning architectures, transformer-based models, sequence models and classical ML where each is appropriate.
• Define training labels, evaluation methods, objective functions, confidence scoring and model performance criteria tied to business outcomes.
• Analyze large structured, time-series, event and text datasets to identify useful features, patterns, causal indicators and failure modes.
• Collaborate with data engineers to create reliable training datasets, feature definitions and reproducible model pipelines.
• Collaborate with backend and platform engineers to move models from experiment into production inference and monitoring.
• Perform error analysis, drift analysis, root-cause investigation and iterative model improvement.
• Communicate model behavior, assumptions, limitations and trade-offs clearly to technical and non-technical stakeholders.
• Use generative models, embeddings, retrieval methods or LLM workflows when they improve the business decision system.
• Create practical documentation so models can be maintained, audited and improved over time.

Requirements and Skills

• 6+ years of experience in data science, applied ML, AI research engineering or related roles working with large datasets.
• Strong foundation in machine learning, statistics, probability, optimization, model evaluation and experimental design.
• Hands-on experience training and evaluating neural networks, deep learning models, sequence models or transformer-based models.
• Strong Python skills and experience with ML tools such as PyTorch, TensorFlow, JAX, scikit-learn, XGBoost, MLflow, Hugging Face or similar frameworks.
• Experience with forecasting, anomaly detection, classification, regression, ranking, recommendation, clustering or root-cause analysis methods.
• Experience working with big data tools such as Spark, Databricks, distributed data processing, large SQL datasets or cloud-scale data platforms.
• Ability to define meaningful metrics that connect model accuracy and model behavior to business optimization outcomes.
• Experience collaborating with engineers to productionize models, not only creating offline analyses.
• Strong communication skills and ability to explain complex model results in clear language.
• Strong academic background in Computer Science, AI, Data Science, Statistics, Applied Mathematics, Engineering or equivalent experience.

Preferred Background

• Experience with LLMs, embeddings, RAG, fine-tuning, prompt evaluation, model distillation, model compression or agentic AI workflows.
• Experience with time-series foundation models, transformers for structured data or high-scale forecasting systems.
• Experience with feature stores, model monitoring, drift detection, explainability tools and MLOps workflows.
• Experience in business optimization problems where model outputs influence operational decisions, prioritization or resource allocation.

Startup Environment

This is a startup environment for people who want meaningful responsibility rather than narrowly defined corporate roles. Team members should expect exposure to multiple parts of the business, including product design, engineering decisions, customer problem solving, implementation planning and operational execution. The team will be small, highly technical and organized around talented builders who can work directly with one another without unnecessary layers of hierarchy. We expect people to use modern development tools aggressively, including coding assistants, automation, test tools, model tooling and sufficient token budgets where they improve speed and quality. The working style is flexible, but the expectations are high: clear ownership, written thinking, disciplined execution, frequent communication, clean handoffs and the ability to make progress without waiting for a complete corporate structure.

What Success Looks Like

• Models move from exploration to validated production use with clear metrics and monitoring.
• Model outputs improve decision quality, prioritization, forecasting or business optimization outcomes.
• The data science function becomes rigorous, reproducible and deeply connected to engineering execution.