Product Analytics Solution Architect

Overview

The Product Analytics Solution Architect is the end-to-end technical owner of an enterprise analytics platform spanning a governed analytics function library, visual workflow builder, agentic AI analyst, and tool/integration layer. The role evolves into an AI Solutions Architect, designing, reviewing, and governing agent, retrieval, evaluation, and safety architecture in line with enterprise architecture, security, and Responsible AI standards.

What you'll do

• Own the end-to-end architecture from data sources through the analytics library, tool/integration layer, AI agent, and workflow builder.
• Define scalable, secure, reusable patterns for APIs, agent tools, event flows, and platform extensions.
• Govern API lifecycle, including versioning, backward compatibility, deprecation, and contract testing.
• Own non-functional requirements including latency budgets, throughput, availability/SLOs, capacity planning, and resilience.
• Govern integrations with enterprise data sources and ongoing data-warehouse migrations.
• Maintain the architecture roadmap, reference architectures, and Architecture Decision Records; drive build-versus-buy and vendor/tooling evaluations and manage technical debt.
• Design RAG and knowledge architecture covering ingestion, chunking, embeddings, vector/hybrid search, retrieval quality, and freshness.
• Own the semantic layer, data contracts, and metadata/catalog for mapping natural language to governed data.
• Define agent memory and context architecture, including session state, short- versus long-term memory, context-window optimization, and semantic caching.
• Architect the agent backbone, including orchestration/planner-executor patterns, tool boundaries, memory, evaluation hooks, guardrails, and multi-agent patterns as needed.
• Define the LLM provider strategy and cost/FinOps architecture, including model routing, caching, token economics, and budget guardrails.
• Own LLMOps lifecycle architecture, including model/prompt versioning and registries, canary/A-B prompt rollouts, and model upgrade and deprecation strategy.
• Own evaluation and observability architecture, including LLM tracing, trace stores, golden datasets, eval-in-CI, online/offline evaluation, drift, and regression gates.
• Establish AI governance practices, including design reviews, risk assessments, Responsible AI alignment, and agent and prompt lifecycle management.
• Own the AI threat model covering prompt injection, data exfiltration via tools, jailbreaks, output sanitization, and least-privilege tool authorization.
• Architect multi-tenancy, data scoping, row-level security, PII/DLP, and audit/traceability across data and prompts.
• Ensure alignment with enterprise architecture, IT controls, confidential-data handling, and emerging AI regulatory requirements.
• Lead sprint planning, code reviews, technical design documentation, coding standards, and release cadence.
• Stay hands-on 30–50% by building reference implementations, unblocking complex tickets, and prototyping risky areas.
• Own end-to-end troubleshooting across platform components and integration points.
• Mentor engineers on architecture principles, integration patterns, and AI solution design; contribute to hiring, interviewing, team topology, and onboarding/enablement.
• Advise the PM, sponsor, AI Center of Excellence, enterprise IT, and security partners.
• Lead or participate in architecture review boards, design-governance forums, and change-impact assessments.
• Define and report architecture success metrics, including reliability, latency, cost-per-insight, evaluation/accuracy, and adoption readiness.
• Communicate complex architectural concepts and trade-offs to technical and business audiences.

What you'll need

• Bachelor's or Master's in Computer Science, Engineering, or a related field.
• 8+ years building production software.
• At least 3 years as a tech lead or solutions architect on a data, analytics, or AI platform.
• Ability to convert business requirements into clean, modular, well-documented technical designs and reference architectures.
• Deep knowledge of integrating heterogeneous applications, databases, and platforms, with disciplined API/contract governance.
• Strong grounding in distributed systems, asynchronous patterns, scalability, observability, resilience, and secure-by-design.
• Ability to anticipate, diagnose, and resolve complex multi-system issues, including non-deterministic AI failure modes.
• Ability to advise product, engineering, and senior leadership on technology trade-offs and risk.
• Expert-level Python.
• Working knowledge of TypeScript / Node.js for full-stack reviews.
• Production experience on AWS and/or Azure, including compute, storage, networking, and IAM.
• Experience with cloud data warehouses such as Snowflake, BigQuery, or Databricks.
• Strong SQL and experience with data modeling, vector stores, and retrieval.
• Hands-on experience with LLMs, agent frameworks including Model Context Protocol, function-calling, LangChain, LlamaIndex, Semantic Kernel, or Agent SDKs, RAG, context/memory design, and evaluation harnesses.
• Experience with model gateways, prompt/agent registries, LLM tracing, evaluation pipelines, and drift detection.
• Experience with CI/CD pipelines, Git workflows, containerization, and IaC concepts.
• Secure-by-design mindset, AI-specific threat modeling, authorization/least-privilege, and enterprise confidential-data handling.

Nice to have

• Experience designing or governing production AI agents, copilots, or AI-enabled enterprise applications at scale.
• Strong grasp of Responsible AI, model governance, ethical AI, and emerging AI regulation.
• Prior architecture role on a multi-tenant analytics platform or developer platform.
• FinOps for LLM/compute, including optimizing cost-per-query and inference spend at scale.
• Experience integrating with industrial / engineering data ecosystems.
• Recognized open-source or internal-tools contributions and architecture thought leadership.

Details

• Location: Chennai, Tamil Nadu.