Principal AI/ML Engineer Lead

Lead AI/ML engineering standards and shared agent infrastructure for Trintech’s AI Financial Close solutions. This full-time, hybrid role is based in India - Bangalore and works across Agent Stream pods and the AI Platform team.

Responsibilities

• Define platform-wide conventions for agent frameworks, prompt lifecycles, evaluation harnesses, guardrails and confidence threshold calibration. Own Langfuse instrumentation, trace validation, evaluation pipelines, prompt regression testing and model-version regression detection.
• Standardise RAG architecture for financial document reasoning, including embeddings, vector databases, retrieval, reranking and structured outputs. Review pod-level agent designs for prompt quality, memory architecture, evaluation rigour and production reliability; contribute to AI/ML hiring standards and interviews.
• Design the platform’s personalised, tiered agent memory architecture: working, episodic, semantic and procedural memory. Select appropriate frameworks such as LangMem, Mem0, Zep/Graphiti or Letta; establish extraction, deduplication, contradiction-resolution and forgetting policies. Define context management for long-running, multi-step financial close workflows.
• Work with the Platform Architect to expose RAG, memory and evaluation infrastructure as shared services. Identify cross-pod quality gaps and assess emerging frameworks, memory systems, evaluation approaches and protocols such as MCP and A2A for potential adoption.

Requirements

• Extensive experience in a relevant technical discipline and delivering production-grade AI or ML solutions; hands-on production LLM agent development using LangChain, LangGraph or similar frameworks. Experience setting engineering standards across teams, designing RAG and agent memory systems, and managing prompt versions, rollback, evaluation and regression testing.
• Knowledge of LLM observability, guardrails, model evaluation, statistical reasoning and experimental design. Production Python and API-framework experience, such as FastAPI, plus PostgreSQL, pgvector or similar vector technology, Docker, Kubernetes and Azure OpenAI or an equivalent provider. Strong technical leadership, communication, collaboration and problem-solving skills.

Preferred

Experience with Ragas or equivalent RAG evaluation frameworks; model fine-tuning or RLHF; and financial close, Record-to-Report, accounting or enterprise finance software.