Product Architect

Fractal is a leading AI and analytics company helping global enterprises turn data into decisions at scale. We are hiring a Product Architect to own the end-to-end technical architecture across our portfolio of Agentic AI platforms and Data platforms. This is a hands-on architecture role that combines deep technical ownership with product thinking, spanning front end, back end, agent orchestration, data, cloud, and DevSecOps.

This role is ideal for a seasoned architect who can define and evolve the architecture of multi-tenant, multi-LLM, multi-framework agentic products and enterprise data platforms used by global clients. You will make key architecture decisions, set engineering standards, design for scale and extensibility, and act as the technical anchor across product, engineering, and client engagements.

You will work closely with product managers, engineering leads, data engineers, AI/ML engineers, and client-facing teams to translate roadmap priorities into robust, governable, and extensible architecture that works across multiple LLM providers (e.g., Claude, OpenAI, Gemini, open-source models) and agent frameworks (e.g., CrewAI, LangGraph, AutoGen, ADK).

Key Responsibilities

- Own the end-to-end architecture of agentic products across their surfaces: low-code/canvas workflow designers, conversational agent interfaces, CLI, IDE extensions, REST APIs, MCP servers, and web interfaces.
- Architect core abstraction layers: agent definition and orchestration (framework-agnostic), LLM abstraction and multi-model routing, and governance/observability across agents deployed org-wide.
- Design scalable, secure, multi-tenant architecture for enterprise deployment, including data isolation, RBAC, audit trails, and compliance controls appropriate for regulated client environments.
- Define architecture for agent-to-agent and agent-to-tool interoperability, including MCP server design, tool/connector patterns, and emerging agent interoperability protocols (e.g., A2A, ACP).
- Lead technical design for data platform capabilities: data pipelines, migrations, data quality, and semantic/knowledge layers (including knowledge graphs and ontologies) that ground AI and analytics on trustworthy, well-modeled data.
- Set and enforce architecture standards, design patterns, and documentation practices (architecture diagrams, ADRs, API contracts) across a distributed engineering organization.
- Evaluate build-vs-integrate decisions against the evolving agentic ecosystem (agent harnesses, session layers, orchestration frameworks), ensuring our products maintain a defensible, non-commoditized architectural position.
- Define the evaluation and testing strategy for agentic systems, including eval harnesses, regression testing for prompts/agents, and structured red-teaming.
- Own cost governance for LLM/agent usage: budgeting,



tenant-level chargeback, token/latency monitoring, and cost-optimization across model providers.
- Establish versioning and change-management practices for agents and prompts, so updates don't silently break downstream consumers.
- Define human-in-the-loop and approval-workflow patterns for higher-risk agentic actions, balancing autonomy with enterprise trust and control.
- Partner with client-facing teams on architecture reviews, PoC/RFP technical responses, and solution estimation for enterprise engagements.
- Drive non-functional excellence: performance at scale, reliability, observability, cost/latency optimization across LLM calls, and secure SDLC/DevSecOps practices.
- Mentor senior engineers and product engineers, conduct architecture reviews, and raise the technical bar across the organization.
- Stay ahead of the rapid-moving agentic AI landscape and translate emerging patterns (agent memory, evaluation frameworks, multi-agent orchestration, knowledge graphs/ontologies for grounding) into pragmatic architecture decisions.

Must Have

- 12-18+ years of experience building enterprise-grade software platforms, including significant time in architecture or senior technical leadership roles.
- Deep hands-on experience with at least one hyperscaler (AWS, Azure, or GCP) at an architecture level - networking, IAM, managed compute/storage, and well-architected framework principles.
- Experience with Infrastructure-as-Code and GitOps-style infra delivery.
- Experience designing for high availability and disaster recovery in production cloud environments.
- Strong hands-on expertise across the full stack: JavaScript/TypeScript, HTML5, CSS3, React/Next.js on the front end; Python with FastAPI and/or Django on the back end.
- Proven experience architecting distributed, multi-tenant, API-driven systems at enterprise scale, with strong grasp of system design, security, and performance trade-offs.
- Hands-on experience with agentic AI systems: LLM orchestration, agent frameworks (e.g., LangGraph, CrewAI, AutoGen, ADK), prompt orchestration, tool use, and multi-agent workflows.
- Working knowledge of the Model Context Protocol (MCP) or comparable agent-tool/agent-agent interoperability standards.
- Strong command of design patterns, architecture documentation (diagrams, ADRs), and the ability to communicate trade-offs clearly to technical and non-technical stakeholders.
- Experience with relational and NoSQL databases, data modeling,



and query optimization.
- Solid DevOps/DevSecOps fundamentals: Docker, Kubernetes, CI/CD pipelines, cloud-native deployment (AWS/Azure/GCP), observability, and secure SDLC practices.
- Experience with REST APIs and familiarity with event-driven and streaming integration patterns (webhooks, message queues, WebSockets/SSE).
- Strong understanding of AI architecture considerations: context/session management, cost and latency trade-offs, guardrails, evaluation, monitoring, and responsible AI controls.
- Excellent stakeholder management and communication skills, with experience operating in client-facing or pre-sales technical contexts.

Nice to Have

- Direct experience building or architecting agentic platforms, agent orchestration frameworks, or developer tooling (CLI, IDE extensions, low-code canvases).
- Multi-cloud or hybrid-cloud architecture experience.
- Cloud certifications (e.g., AWS Solutions Architect Professional, Azure Solutions Architect Expert, GCP Professional Cloud Architect).
- Experience with cloud-native managed AI/ML platforms (Bedrock, Vertex AI, SageMaker, Azure AI Foundry) as build-vs-buy alternatives to self-hosted inference.
- Hands-on knowledge and experience of Agentic AI development tools such as Cursor, Windsurf, and Claude Code, including using them productively in day-to-day engineering.
- Working knowledge of knowledge graphs, ontologies, and semantic data modeling, and how they can ground and improve agentic/AI system reliability.
- Exposure to vector databases, RAG architectures, agent memory strategies, and GenAI evaluation frameworks in production.
- Experience with modern data platform ecosystems (e.g., Databricks, Snowflake, or equivalent lakehouse/warehouse platforms).
- Experience contributing to RFP responses, technical proposals, or reusable accelerator IP in a consulting or platform product context.
- Familiarity with multi-LLM routing/gateway patterns and cost/performance optimization across model providers.
- Prior experience in a "framework-and-model-agnostic" or "no vendor lock-in" platform positioning.

What Success Looks Like

- Architecture scales cleanly across new LLM engines, agent frameworks, and enterprise clients without accumulating technical debt.
- Clear, well-documented architecture (diagrams, ADRs, API contracts) that the engineering organization can build against confidently.
- Defensible architectural differentiation versus adjacent tools (agent frameworks, low-code automation platforms, session/harness layers) grounded in governance, extensibility, and organizational persistence rather than feature checklists.
- Strong client confidence in technical reviews, PoCs, and RFP engagements.
- A mentored, technically strong engineering team operating against consistent architecture standards.

📌 Product Architect (Bengaluru)
🏢 Fractal Analytics
📍 Bengaluru