Data Science engnieer

We are hiring a senior individual contributor to join our Data Science team. This is a hands-on role for someone who

can take ownership of the full solution, from understanding the problem and building the required data pipelines to

developing the model, deploying it, and keeping it reliable in production.

Core Focus & Algorithmic Development

• Custom Model Engineering: Design, train, and deploy custom PyTorch architectures from scratch to solve non-standard problems, moving beyond off-the-shelf fine-tuning or tutorials.
• Sensor & Telematics Data Processing: Develop models targeting complex physical, industrial, IoT, or automotive systems handling noisy, delayed, or weakly-labelled data.
• Advanced Predictive Analytics: Implement models for Anomaly Detection, Survival Analysis, Remaining Useful Life (RUL), and Failure Prediction.

Production Ownership & MLOps Infrastructure

• End-to-End Lifecycle Ownership: Drive machine learning models through their complete lifecycle—from initial formulation to active production monitoring, automated retraining pipelines, and live troubleshooting.
• Production-Grade Software Engineering: Write clean, modular Python code backed by automated unit/integration tests and CI/CD pipelines.
• Engineering Best Practices: Own and maintain workflow orchestration, cloud infrastructure, alerting/monitoring systems, and participate in technical on-call rotations.

Distributed Computing & Big Data Engineering

• Distributed Data Processing: Leverage Apache Spark for deep performance tuning, optimal partitioning, memory management, and compute cost optimization.
• Big Data Platform Management: Build and maintain scalable data pipelines on platforms such as Databricks (preferred), Snowflake, or BigQuery.

Production LLMs & Agentic Systems

• Agentic Application Deployment: Ship LLM-driven and agentic applications directly into production environments.
• Operational Guardrails: Implement comprehensive evaluation frameworks (evals), safety guardrails, and latency/cost optimization controls for deployed generative systems.

Must have:

Over 5 years industry experience in a data science or machine learning role

2. Experience taking machine learning to production - Owned in production,

including monitoring, retraining, and troubleshooting (prototyping or only deploying

and handing over to others is not enough)

Python – has to have at least writes tested production modules / Owns production

services with CI/CD (only Notebooks and scripts is not enough)

PyTorch

• Built and trained custom architectures and Shipped custom architectures

to production (Tutorials and fine-tuning is not enough)

Spark and distributed computing

• Regular production use or Deep — tuning,

partitioning, cost optimization (Occasional is not enough)

6. Big data platforms – must have at least one – Databricks, Snowflake, BigQuery,

Hadoop/Hive (preferably Databricks)

Must Owen one or more production engineering practices

• Automated testing,

CI/CD, Workflow orchestration’ Monitoring & alerting, cloud infrastructure, on-call

Preferred

8. MSc in Computer Science or an engineering-related quantitative field, such as Electrical

Engineering, Mechanical Engineering, Statistics, Physics, or Applied Mathematics. A PhD is an

advantage.

9. Worked with on noisy or weakly-labelled data such as - Anomaly detection,

Survival analysis, Remaining useful life, Failure prediction with delayed labels

10. Domain experience - Automotive/telematics, IoT sensors, Predictive maintenance,

Industrial or physical systems

LLM or agentic applications to production

• Shipped or Shipped with evals,

guardrails, and cost/latency management (only experimented is not enough)

Shipped or Shipped with evals, guardrails, and cost/latency management –

Advantage, experimented can work as well.

Please note – one with none of the preferred is not relevant – should be balanced in the preferred

(meaning you can miss on one or two from the preferred but not more).