Bengaluru, Karnataka, India
Remote only
This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for an AI/ML Solution Architect based in India.
This is a high-impact architecture role focused on designing production-grade AI systems for education, assessment, and other large-scale public-sector environments. You’ll bridge discovery and delivery, translating complex stakeholder requirements into secure, scalable, and cost-conscious technical architectures. The role spans GenAI, RAG, multi-agent systems, document AI, LLM evaluation, cloud platforms, and LLMOps/MLOps. You’ll work closely with engineering teams, programme leaders, domain specialists, and institutional stakeholders across multiple countries. Your designs will need to operate reliably under demanding requirements around privacy, auditability, security, multilingual content, and sovereign hosting. This is an opportunity to shape AI systems intended for real-world deployment and measurable impact at significant scale.
Lead solution discovery workshops with programme, government, domain, and engineering stakeholders, translating complex or incomplete requirements into actionable technical architectures.
Own solution design from initial discovery through production, producing service breakdowns, data flows, trust boundaries, sequence diagrams, architecture decision records, cost models, and implementation-ready blueprints.
Architect production AI pipelines covering generation, RAG, knowledge-graph retrieval, multi-agent workflows, LLM-as-judge systems, document AI/OCR, human review, and constraint-based processing.
Define AI system contracts covering inputs, outputs, schemas, model and prompt versions, token usage, latency, evaluation criteria, and failure-handling behavior.
Design secure human-in-the-loop workflows with appropriate segregation of duties, audit trails, configuration controls, de-identification, and protection of sensitive information.
Own reference architectures across Azure and/or AWS, including identity and access management, private networking, storage, vector search, monitoring, observability, LLMOps/MLOps, and production operations.
Establish evaluation frameworks and release gates that can prevent underperforming models, prompts, or configuration changes from reaching production.
Incorporate FinOps principles into architecture, including token budgets, inference and OCR cost controls, model routing, fallbacks, and degradation strategies.
Remain accountable through implementation and the first production cycle, ensuring architectures are practical, resilient, secure, and operationally viable.
Produce reusable architecture patterns, points of view, technical documentation, and country-adaptation frameworks while mentoring technical leads on the appropriate boundaries between AI-driven and deterministic workflows.
8+ years of experience building software, data, or AI/ML systems, including recent experience as a Solution Architect, AI/ML Architect, Principal Engineer, or Staff Engineer with end-to-end architecture ownership.
Proven experience shipping at least one production AI/ML or GenAI system at scale, such as RAG, multi-agent systems, LLM-as-judge pipelines, or classical ML/NLP solutions that successfully passed security, cost, and operational reviews.
Demonstrated ability to create delivery-ready architecture blueprints covering services, data stores, IAM, trust boundaries, failure modes, evaluation strategies, and cost models.
Deep experience with Azure and/or AWS, including the ability to design and defend landing zones, private-endpoint architectures, cloud security controls, and real-world infrastructure and inference costs.
Strong Python skills, with practical understanding of classical machine learning, model evaluation, LLM APIs, embeddings, vector search, and production AI behavior.
Hands-on experience with LLMOps/MLOps, including model and prompt versioning, evaluation harnesses, CI/CD release gates, observability, experiment tracking, and rollback strategies.
Strong understanding of RAG over structured knowledge, hybrid search, document AI/OCR, multilingual AI, agent security, prompt-injection risks, and AI evaluation methodologies.
Experience working with non-technical stakeholders and facilitating workshops that translate business, programme, or government requirements into technical solutions.
Excellent written English and strong documentation skills, with the ability to create architecture materials that function as clear technical contracts for engineering teams.
Strong problem-solving, decision-making, communication, and collaboration skills, with the ability to work independently while influencing technical direction across multidisciplinary teams.
Preferred experience includes Azure AI Foundry, Azure OpenAI, Azure Document Intelligence, AWS Bedrock or SageMaker, GCP Vertex AI or BigQuery, MLflow, Databricks, Microsoft Fabric, Semantic Kernel or Microsoft Agent Framework, MCP integrations, constraint solvers, and exam or assessment platforms.
Experience in education, assessment, public sector, healthcare, finance, identity, or other high-audit environments is highly valuable, as is familiarity with sovereign hosting, government data systems, multilingual or code-mixed language applications, and data-sharing requirements.
Relevant cloud certifications, such as Azure Solutions Architect or Data Scientist and AWS Solutions Architect or Machine Learning certifications, are advantageous.
The role requires comfort working across international time zones, with early mornings or late evenings occasionally forming part of the working rhythm.
Fully remote-first position based in India, with occasional travel for client or stakeholder workshops.
Full-time opportunity with significant ownership over architecture and production AI systems.
Opportunity to work on AI applications spanning education, assessment, document intelligence, multilingual technology, and public-sector systems.
Meaningful opportunity to contribute to technology designed to support learning and assessment outcomes at large scale.
Exposure to advanced GenAI, RAG, multi-agent, LLM evaluation, OCR, LLMOps/MLOps, cloud architecture, and AI safety practices.
Collaborative international environment with colleagues and stakeholders across multiple regions.
Opportunity to influence technical strategy, reusable architecture patterns, engineering practices, and future AI capabilities.
Competitive remuneration based on experience and skillset.
Flexible remote working environment, with international collaboration across time zones.
Opportunities for professional growth through technically challenging, mission-driven projects and close collaboration with multidisciplinary teams.
