Job Requirements

Job Title: Principal AI Architect (Generative AI & Agentic Systems) Experience Range: 8+ Years (2+ years leading GenAI/Agentic production architectures) Location: Bangalore / Hyderabad / Gurgaon (Hybrid)

Job Overview

We are seeking a seasoned Principal AI Architect to serve as the chief technical authority for our next-generation Generative AI and autonomous agent initiatives across high-throughput, heavily regulated enterprise environments. In this role, you will not simply implement predefined architectures; you will analyze ambiguous, mission-critical business challenges (e.g., real-time fraud arbitration, customer friction reduction, operational queue unbottlenecking) and evaluate technical trade-offs to decide which architecture is required. You will design fit-for-purpose agentic systems—determining when to deploy single-agent co-pilots, hierarchical supervisor-worker topologies, or choreographed multi-agent networks—built on scalable cloud infrastructure (Google Cloud Platform) and powered by frontier reasoning models This role requires an architect who pairs deep algorithmic and model-behavior comprehension with battle-tested enterprise systems design, low-latency data engineering, and rigorous regulatory compliance.

Key Responsibilities

- Architectural Strategy & Topology Decisioning

● Problem-to-Architecture Mapping: Analyze complex business friction points, evaluate trade-offs (reasoning accuracy vs. latency vs. cost), and determine the optimal architectural paradigm (e.g., monolithic LLM workflow vs. hierarchical multi-agent supervisor vs. event-driven worker pools).

● Model Strategy & Routing: Establish model selection frameworks across commercial frontier models (Claude, Gemini, etc.) and specialized enterprise risk/decision engines. Design intelligent routing gateways based on task complexity, inference cost, and SLA thresholds.

● Reference Architectures: Define enterprise-wide blueprints, design patterns, and engineering standards for autonomous agent state management, memory persistence, and inter-agent communication.

- Tooling, Integration & Federated Data Strategy

● Orchestration Frameworks vs. Native Build: Evaluate commercial/open-source orchestration frameworks (LangGraph, LangChain, AutoGen, LlamaIndex) against custom, lightweight microservice implementations, determining when framework abstraction is useful versus when native code is required to meet strict latency SLAs.

● Protocol & Tool Standardization:



Evaluate and implement up-to-date agent-tool interoperability standards (such as Model Context Protocol - MCP) to cleanly separate model reasoning from backend transactional execution.

● Federated Zero-Copy Ingestion: Architect real-time, read-only data access layers across heterogeneous enterprise platforms (payments, auth, CRM, credit, and dispute systems) without duplicating data into fragile, monolithic data lakes.

● High-Throughput / Low-Latency Design: Design parallel fan-in/fan-out query pipelines, asynchronous event streaming, and in-memory feature caches to meet mission-critical latency SLAs (sub-second and sub-50ms execution paths).

- Governance, Zero-Trust Security & Regulatory Compliance

● Policy-Enforced Access Control: Architect Attribute-Based Access Control (ABAC) and role-specific synthesis layers, guaranteeing that domain agents access only certified summaries or full records according to regulatory governance.

● Compliance Firewalls: Programmatically enforce statutory constraints within regulated industries (e.g., AML anti-tipping-off rules, consumer credit disclosure regulations, and marketing purpose limitations).

● Security & Redaction: Implement automated prompt injection defense, PII scrubbing, tamper-evident audit logging, and customer-managed encryption key (CMEK) management within enterprise VPC perimeters.

- Human-in-the-Loop & Progressive Autonomy

● Calibrated Autonomy Ladders: Design frameworks that enable operational workflows to safely graduate across progressive autonomy stages: from passive Observation and Decision Recommendation (Co-Pilots) to Human-Approved Action, and strictly bounded Autonomous Execution.

● Evaluation & CI/CD Pipelines: Architect automated evaluation frameworks, golden test datasets, and shadow-execution pipelines to validate model reasoning, mitigate hallucination, and benchmark accuracy before production rollout.

Required Qualifications

● Architecture Leadership: 8+ years in distributed systems architecture, enterprise software engineering, and data platforms, with 2+ years architecting and scaling production Generative AI or autonomous agent systems.

● Architectural Decision-Making:



Proven ability to dissect complex, ambiguous operational problems, formulate architectural alternatives, and rigorously defend architectural trade-offs

● Frontier Model Expertise: Deep technical mastery of frontier reasoning models (particularly Anthropic Claude and Google Cloud ecosystem), including advanced prompt caching, function calling, XML structuring, and token optimization.

● Orchestration & Tooling Standards: Hands-on experience deploying and evaluating multi-agent orchestration frameworks (e.g., LangGraph, AutoGen) and modern integration standards (Model Context Protocol - MCP).

● Cloud Infrastructure: Extensive experience designing on Google Cloud Platform (Vertex AI, Cloud Run, Pub/Sub, Apigee, Cloud Spanner/AlloyDB, Memorystore) or comparable enterprise cloud environments.

● Regulated Industry Fluency: Strong background architecting systems within heavily regulated domains (FinTech, Payments, Banking, Insurance, or HealthTech) where compliance, explainability, and data isolation are non-negotiable.

About Company: Straive

Straive is a market leading Content and Data Technology company providing data services, subject matter expertise, & technology solutions to multiple domains.

Data Analytics & Al Solutions, Data Al Powered Operations and Education & Learning form the core pillars of the company’s long-term vision. The company is a specialized solutions provider to business information providers in finance, insurance, legal, real estate, life sciences and logistics. Straive continues to be the leading content services provider to research and education publishers.

Data Analytics & Al Services: Our Data Solutions business has become critical to our client's success. We use technology and Al with human experts-in loop to create data assets that our clients use to power their data products and their end customers' workflows. As our clients expect us to become their future-fit Analytics and Al partner, they look to us for help in building data analytics and Al enterprise capabilities for them.

With a client-base scoping 30 countries worldwide, Straive’s multi-geographical resource pool is strategically located in eight countries - India, Philippines, USA, Nicaragua, Vietnam, United Kingdom, and the company headquarters in Singapore.

Website: https://www.straive.com/ Linkedin

Straive is an equal opportunity employer. We commit to celebrate diversity, equity and inclusion in the workplace.

📌 Architect (Gurugram)
🏢 Straive
📍 Gurugram