Forward Deployed Engineer (FDE)

As a Forward Deployed Engineer – AI/ML, you take Client’s forecasting models and AI capabilities from the lab into live customer environments, and make them work on each customer's real data.

You sit between our Data Science, Engineering, and Professional Services teams and the customer. You connect customer ERP and master data to the client platform, stand up and tune ML forecasting pipelines on Databricks, debug what breaks in production, and turn what you learn in the field into requirements for Product. You are not expected to invent new model architectures. You are expected to understand them well enough to deploy, configure, evaluate, and explain them.

Experience

5+ years

Location

Pune (Hybrid)

Department

Data Science & ML Engineering (customer-facing)

Type

Full-time

KEY RESPONSIBILITIES

Customer deployment and integration

Custom build: Analyze each customer's planning process and data landscape, then design the APIs, data pipelines, and agentic AI workflows needed to fit the client platform to it.

Data integration: Clean, structure, and connect legacy ERP, WMS, and master data sources into the data platform's scalable integration layer so models and planners get consistent, trusted inputs.

Production readiness: Turn internal prototypes into secure, compliant, production-ready deployments inside customer and client cloud environments.

Live troubleshooting: Debug production pipelines and model runs, resolve schema conflicts and data quality issues, fix edge cases (new items, intermittent demand, missing history), and keep nightly and on-demand runs reliable.

Applied data science and ML operations

Forecast model deployment: Configure, train, and deploy production forecasting models (Adaptive ML) for each customer, including per-customer settings for seasonality, price, promotions, and new product introductions.

ML pipelines: Build and maintain the customer-specific parts of end-to-end ML pipelines: data ingestion, feature engineering, training, scoring, and write-back of results. Orchestrate these as scheduled and event-triggered workflows with clear dependencies, retries, backfills, and alerting, so nightly and on-demand runs complete reliably.

Evaluation: Measure forecast accuracy and bias against legacy models and customer baselines, quantify forecast uncertainty using the model's probabilistic outputs, and explain the drivers behind a forecast to planners.

MLOps: Apply model versioning, monitoring, rollback, and automated retraining; contribute to CI/CD for pipelines and models.

Performance: Profile and tune Spark jobs, Delta table writes, and inference latency so forecasts reach planners on time.

AI applications: Deploy and monitor retrieval-augmented generation (RAG) and agentic AI workflows, with appropriate AI observability.

ESSENTIAL QUALIFICATIONS

Education and experience

B.S. or M.S. in Computer Science, Data Science, Statistics, Software Engineering, or a related field.

5+ years of software or data engineering experience, including at least 2 years putting ML models into production.

Data science and ML

Solid grounding in statistical modeling, regression, and time series forecasting.

Proficiency with pandas, NumPy, and scikit-learn for data preparation, feature engineering, and model evaluation.

Working knowledge of PyTorch or a similar deep learning framework: enough to read, configure, train, and debug existing models.

Understanding of probabilistic forecasts and prediction intervals, and how to communicate uncertainty to business users.

Hands-on MLOps experience: model versioning, monitoring, retraining pipelines, and CI/CD.

Experience on Databricks or a comparable cloud data and ML platform with large datasets.

Hands-on experience deploying ML models and RAG systems, and using AI observability tools.

Customer skills

High emotional intelligence and customer empathy; able to explain a model's output to a demand planner and a pipeline failure to an IT lead.

Comfortable working in ambiguous, unstructured environments and owning a problem through to resolution.

PREFERRED QUALIFICATIONS

Experience in demand forecasting, inventory planning, or retail and distribution supply chains.

Familiarity with interpretable forecasting methods that decompose demand into trend, seasonality, price, and promotion effects.

Experience tuning Delta Lake tables (merges, partitioning, clustering) and Spark jobs at hundreds of millions of rows.

Experience serving models for low-latency inference (model serving endpoints, ONNX, or containerized services).

Cloud deployment experience on AWS, GCP, or Azure, including containerization with Docker and Kubernetes.

Hands-on orchestration with Airflow, Prefect, or Databricks Workflows: building DAGs, managing dependencies and retries, event- and schedule-based triggers, backfills, and alerting on failed runs.

Exposure to explainable AI frameworks and AI observability in production.

Prior forward deployed, solutions engineering, or professional services experience with enterprise SaaS customers.

WHY THIS ROLE

Put modern ML forecasting into production at many distributors, retailers, and manufacturers, and see its impact on their inventory and service levels.

Work across Data Science, Engineering, and Product, with a direct line to how the platform evolves.

A growth path into ML Engineering, Solution Architecture, or Product roles.

Pay: ₹125,000.00 per month

Benefits:

• Work from home

Application Question(s):

• This is a contract role for 3 months and may get extended if needed. are you okay with this?
• job location is pune (shivajinagar)-hybrid model. are you okay with this
• the salary we are providing is 1,25,000 fixed (payroll from sourcedear) with no PF or ESI due to contract position. are you okay with this?
• how soon you can join?

Work Location: Hybrid remote in Shivajinagar, Pune, Maharashtra