AI Principle Architect
Mastercard2 days ago
Pune, IndiaStaff+
Responsibilities
- Own the architecture and technical roadmap for the Operational Intelligence Gen AI platform, including reference architectures and reusable standards.
- Architect enterprise tool-calling systems with schemas, contracts, versioning, access control, retries, timeouts, idempotency, circuit breakers, and fallbacks.
- Define architectures for multi-agent systems, skills-based agents, self-correcting agents, memory, state management, and grounded retrieval.
- Design high-throughput inference and orchestration covering asynchronous processing, streaming, caching, rate limiting, and model routing.
- Establish Databricks-based LLMOps, evaluation, observability, tracing, model lineage, and prompt and agent delivery standards.
- Build security and Responsible AI into the platform, including guardrails, prompt-injection defenses, hallucination mitigation, privacy, and audit trails.
- Own architecture decisions involving build-versus-buy, model and vendor selection, cost, latency, accuracy, and emerging standards such as MCP.
- Lead design reviews, align technical stakeholders, mentor architects and senior engineers, and build hands-on prototypes and reference implementations.
Requirements
- Bachelor’s or Master’s degree in Computer Science, AI/ML, or Engineering, or equivalent practical experience.
- Significant production experience architecting large-scale distributed systems, including microservices, event-driven systems, data platforms, and cloud-native infrastructure.
- Deep expertise in API design and systems integration, including contracts, security models, and reliability patterns.
- Hands-on experience designing and shipping production LLM-powered systems, especially agents using tool or function calling, memory, and retrieval.
- Strong Python skills for prototyping, including asynchronous programming and FastAPI or similar, plus SQL proficiency.
- Deep experience with AWS, Azure, or GCP and related AI services such as Bedrock, SageMaker, Azure OpenAI, or Vertex AI.
- Grounding in enterprise security, privacy, and compliance, preferably in financial services or another regulated industry.
- Ability to influence without authority and communicate technical trade-offs to engineering, product, leadership, and business stakeholders.
- Preferred experience includes tool critics, Self-RAG, CRAG, reflection loops, Mem0, Zep, MCP tool servers, agent evaluation frameworks, fine-tuning with LoRA or PEFT, RLHF, multimodal models, or knowledge graphs.
Tech Stack
Categories
About Mastercard
Mastercard builds and operates a global payments network used by banks, merchants, fintechs, and governments, offering card processing, real-time payments, tokenization, and fraud/risk services. It generates revenue from transaction processing and assessment/service fees across more than 200 countries and territories. Founded in 1966 and headquartered in Purchase, New York, Mastercard is a public company listed on the NYSE (ticker: MA).