Location Name: Pune Corporate Office - Mantri
Job Purpose
As an AI Engineer in Data Intelligence Unit, you will help build and operate the core building blocks of Data Representation learning and Context Engineering across the credit Risk, Fraud/FRM, Sales, Collections & Recovery. You will work with senior engineers and Data scientists to converts raw structured and unstructured data into reliable features, embeddings, retrieval-ready knowledge assets, and repeatable evaluation pipelines so downstream AI pods can ship models faster, safer, and with measurable quality.
Duties And Responsibilities
- Data Representation Pipelines
- Prepare and validate datasets from multiple sources (transactions, bureau, device/digital, documents,
CRM/operations)
- Implement features engineering pipelines (aggregations, ratios, behavior signals) and maintain feature
definitions.
- Build large-scale ML systems: distributed training pipelines, feature stores, model registry, CI/CD for ML, and
scalable batch + near-real-time scoring services.
- Support embedding workflows (text/customer/device/dealer/geo) including batch refresh, versioning, and
lineage.
- Knowledge Engineering Support (Canonical Objects & Metadata Assets)
- Help create/maintain canonical objects, entity dictionaries, taxonomies/ontologies, and labeling guidelines.
- Support annotation/labeling workflows (quality checks, consistency, sampling) for training and evaluation.
- Experimentation & Model Operations
- Execute training/inference jobs using established frameworks, log experiments and outcomes.
- Perform error analysis, data leakage checks, and basic model monitoring (drift signals, data anomalies)
- Contribute to deployment readiness: tests, reproducible configs, and incident triage support.
- Retrieval & Context Engineering Support (LLM/RAG enablement)
- Assist document processing: chunking, cleaning, metadata tagging, indexing access filters.
- Maintain prompt/context templates, grounding rules, and evaluation sets for RAG/LLM assistants used by
Pods.
- Run offline evaluations (retrieval quality, answer quality, regressions) and track metrics across releases.
- Engineering Hygiene & Governance
- Write clean, testable code; follow Git workflows and CI checks.
- Maintain documentation: dataset cards, feature notes, pipeline SOPs, and release checklists.
- Follow security/privacy controls for regulated data, ensuring traceability and auditability.
Basic Qualifications
Required Qualifications and Experience
- Bachelor's/Master's in CS/Math/Engineering
- 0 2 years experience in Data Science /Applied ML/ ML Engineering with proven leadership delivering
production grade ML system at scale.
Required Skills & Competencies Core (must-have)
- Programming: Python (strong), SQL (strong); Git; basic unit testing.
- Data: Pandas/PySpark basics, joins/aggregations/window functions, data validation and profiling.
- ML Fundamentals: supervised/unsupervised learning, embeddings, train/val/test discipline, metrics, and error analysis.
- Applied System Mindset: reproducibility, structured debugging, logging/monitoring fundamentals. Good-to-have Skills
- ML frameworks: Pytorch / TensorFlow; experiment tracking (MLFlow)
- Retrieval stack: vector indexing concepts, chunking strategies, hybrid search ideas, evaluation datasets.
- Data/Infra: Airflow/Prefect, Spark, Elasticsearch/OpenSearch, MongoDB, feature stores, graph Database, vector database, model serving basics. Preferred Qualifications
- Experience building end-to-end decisioning platforms with real-time and batch orchestration.
- Graph ML / entity-resolution experience for relationship-based risk and fraud analysis.
- Experience operating ML systems across multiple products/segments with multi-tenant controls.
- Publications/patents or strong track record of innovation in applied ML, large-scale ML systems
