5 hours ago
Mumbai, IndiaSenior
Responsibilities
- Own end-to-end AI solution architecture, including system designs, service boundaries, data contracts, implementation plans, and build-versus-buy decisions.
- Design RAG systems with vector databases, hybrid retrieval, metadata filtering, reranking, document parsing, chunking, embeddings, and indexing.
- Evaluate and fine-tune models, curate datasets, implement supervised fine-tuning and parameter-efficient techniques, and benchmark quality, safety, latency, and cost.
- Build Python and TypeScript services, APIs, asynchronous workflows, enterprise data integrations, ingestion pipelines, normalization, schema validation, lineage, and recovery processes.
- Develop stateful enterprise agentic systems with planning, tool use, memory, structured outputs, multi-agent orchestration, MCP integrations, scoped permissions, and human approval controls.
- Implement AI security and governance controls for prompt injection, sensitive-data leakage, unauthorized tool execution, tenant isolation, data residency, retention, and auditability.
- Define AI evaluation and observability practices covering retrieval quality, groundedness, task completion, tool correctness, latency, token usage, and cost.
- Lead architecture and code reviews, mentor AI Specialists, drive root-cause analysis, and improve developer productivity through reusable libraries, reference architectures, tooling, and runbooks.
Requirements
- 6+ years of experience in applied AI, machine learning, or software development, including architecture and production ownership of complex AI systems.
- Experience with enterprise AI controls such as scoped tool authorization, tenant isolation, PII handling, prompt-injection defenses, audit logging, and human approval for sensitive actions.
- Experience designing enterprise semantic layers or context graphs connecting business entities, relationships, and definitions across systems.
- Strong experience with GCP or AWS cloud deployments, serverless technologies, GitHub CI/CD pipelines, Docker, Kubernetes, and model platforms such as Vertex AI or Bedrock.
- Familiarity with centralized LLM gateways such as LiteLLM, including routing, fallbacks, rate limits, token budgets, caching, and access policies.
- Production-grade experience with Python, TypeScript, and SQL, including typed interfaces, APIs, asynchronous processing, and automated testing.
- Experience building enterprise agentic systems with LangGraph, Mastra, or equivalent frameworks, including orchestration, state and memory management, checkpointing, and failure recovery.
- Applied knowledge of LLM inference and context engineering, including structured outputs, embeddings, hybrid retrieval, reranking, and evaluation-driven model selection or fine-tuning.
- Practical experience with Databricks, Spark, and Postgres across analytical and transactional workloads, batch and real-time pipelines, and messy customer-data normalization.
- Experience with Braintrust or equivalent AI evaluation systems, including golden datasets, custom scorers, regression testing, and calibrated LLM-as-judge workflows.
- Experience with enterprise AI observability layers covering agent traces, model and tool spans, prompt versions, error attribution, latency, token usage, and cost tracking.
Tech Stack
Categories
About Coursera
Coursera builds an online learning platform for individuals, enterprises, and academic institutions, offering courses, Professional Certificates, and online degrees from university and industry partners. Its revenue comes from paid courses and subscriptions, enterprise contracts, and tuition-sharing for degree programs. Founded in 2012 by Andrew Ng and Daphne Koller and headquartered in Mountain View, it is publicly traded on the NYSE (COUR). Coursera has combined with Udemy to form a single, larger skills platform.
