Job Description: Job Summary:
We are looking for a senior AI Engineer with 9–12 years of experience to join our Digital and AI Solutions Engineering team, focused on building a next-generation agentic platform for Finance and Accounting operations. You will be a core contributor to designing and delivering platform-level AI capabilities — including multi-agent orchestration, human-in-the-loop workflows, intelligent exception management, and shared intelligence services — that power automated end-to-end processes across Purchase-to-Pay, Order-to-Cash, and Record-to-Report domains.
This is a platform engineering role, not a one-off use case role. You will be building reusable, production-grade services that connect to enterprise ERP systems, process high volumes of financial documents, apply configurable business rules, and route exceptions intelligently — all while maintaining the audit trail and compliance standards that regulated finance environments demand. You will work at the intersection of AI engineering, enterprise integration, and finance operations, collaborating closely with domain SMEs, integration engineers, and delivery teams to ensure that what gets built is both technically sound and operationally adoptable by clients.
The ideal candidate combines strong hands-on experience in LLM-based application development and multi-agent frameworks with the software engineering discipline to build scalable, multi-tenant platform components. Experience in Finance and Accounting processes or prior delivery in a BPO or shared services context is a significant advantage, as the role requires translating complex finance workflows and business rules into robust, configurable AI-driven automation.
Must have Skills:
6+ years of hands-on experience building multi-agent systems using agentic frameworks (LangGraph, AutoGen, or CrewAI) — not just single-chain LLM applications
Experience with event-driven architectures and async workflow patterns (Kafka, AWS SQS, or equivalent) for long-running, stateful processes
Experience integrating AI applications with enterprise APIs — REST, SOAP, or ERP systems (SAP, Oracle, or D365)
Production MLOps experience — model monitoring, drift detection, and evaluation pipelines, not just model building
Experience with document intelligence or structured extraction from unstructured documents (invoices, PDFs, emails)
Experience building human-in-the-loop (HITL) workflows where AI and human decisions are interleaved
Experience with workflow orchestration tools (Apache Airflow, Temporal, or AWS Step Functions)
Familiarity with cloud platforms (AWS/GCP/Azure) and MLOps best practices.
Excellent communication skills with the ability to translate technical solutions into business impact.
Good to have skills:
Familiarity with multi-tenant platform design and configuration-driven architecture (serving multiple clients without code changes per client)
Knowledge of financial controls, audit trail requirements, or SOX-scoped processes
Experience with OCR and document AI services (AWS Textract, Azure Document Intelligence, Google Document AI)
Exposure to ERP-specific integration patterns — SAP RFC/BAPI, Oracle Fusion REST APIs, or D365 connectors
Understanding of anomaly detection in transactional data (not just text/NLP use cases)
Key Responsibilities:
Design and implement multi-agent orchestration workflows with stateful execution, conditional branching, human-in-the-loop handoffs, and retry logic
Build and maintain ERP and enterprise system connectors (SAP, Oracle, D365) as part of a reusable integration layer — ensuring agents can read from and write back to source systems
Develop exception detection, classification, and routing logic within agentic pipelines — including upstream prevention rules and lifecycle tracking
Instrument AI services in production — logging, tracing, latency monitoring, and drift detection across live agent workflows
Design shared platform services (data extraction, anomaly detection, NLP classification, duplicate detection) as reusable, independently deployable microservices consumed across multiple workflows
Collaborate with finance domain SMEs to translate P2P, O2C, and R2R business rules into configurable validation logic — without hardcoding client-specific behaviour
Contribute to a multi-tenant platform architecture where new client onboarding is achieved through configuration, not custom development Evaluate model performance, mitigate bias, and optimize accuracy, latency, and cost.
Stay up to date with the latest trends in LLMs, transformers, and GenAI architecture.
Preferred Qualifications:
Prior experience delivering AI solutions in a Finance & Accounting BPO or shared services environment
Exposure to SOX compliance, internal audit requirements, or financial controls in a technology context
Experience building configurable, metadata-driven integration pipelines with canonical data models
Familiarity with human-in-the-loop (HITL) design patterns — task queuing, approval routing, SLA escalation, and decision capture
Hands-on experience with document AI services for financial document processing (invoices, remittances, purchase orders)
Knowledge of workflow state management patterns — checkpointing, idempotency, and safe resume for long-running processes
Experience with vector similarity search applied to duplicate detection or semantic matching in transactional data
Responsibilities: Lead the development and optimization of complex data models, Oversee data science project management and execution, Mentor junior data scientists, Ensure data solutions align with organizational goals, Provide strategic guidance on data science technologies.
Qualifications: Bachelor's/Master's in Engineering 5-8 years