New opportunity

Principal Engineer, AI Platform

Accrete · Mumbai, India

About this role

Accrete seeks a hands-on Principal Engineer for its AI Platform team in Andheri East, Mumbai. The role leads the architecture, development and scaling of the platform behind the company’s AI products and agentic systems, turning AI capabilities into reliable production services. The position offers hybrid work, with remote flexibility subject to company guidelines.

Responsibilities

Set technical direction across frontend, backend, infrastructure and platform services. Design secure, maintainable systems for AI agent workloads, balancing scalability, performance, reliability, cost and delivery time.
Build end-to-end features, APIs, services, core libraries and developer-facing interfaces for working with AI models, agents and data. Develop infrastructure to securely run and orchestrate agents in production, and integrate AI/ML models and data pipelines.
Identify architectural risks and improve availability, fault tolerance, latency, resource use, observability and developer experience. Establish practices for design, code quality, testing and security; guide the platform roadmap and assess emerging AI/ML technologies.
Own initiatives from architecture through deployment and iteration. Lead design and code reviews, mentor engineers, and collaborate with product, ML, data and engineering teams.

Requirements

A PhD in Computer Science, AI, ML or a related field, or a master’s degree in a relevant technical field from a Tier I institution. The posting’s opening also states “Ph.D Required.”
At least 10 years of software engineering experience, including scalable production systems and platforms; experience owning complex initiatives and providing technical leadership.
Strong frontend experience with React, TypeScript or similar technologies; backend experience with Python, Go or similar languages; knowledge of REST, GraphQL, databases, cloud platforms (AWS, GCP or Azure), Docker, Kubernetes, CI/CD, DevOps and production observability. Familiarity with AI/ML frameworks, model-serving infrastructure or AI application architectures is expected. Experience with AI/ML systems, agents, language models or data-intensive applications is highly desirable.

Preferred

Experience with agent platforms and orchestration, LLMs or generative AI, PyTorch or TensorFlow, real-time or event-driven systems, Apache Spark, secure execution environments, developer platforms, or security, compliance, privacy and government-focused work.

Benefits include flexible PTO, ₹10 lakh family mediclaim coverage, professional development, catered onsite meals and team activities. Salary is described as competitive but no amount is given.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

AI platform architectureAI agentsAgent orchestrationSoftware architectureSystem designDistributed systemsFull-stack developmentReactTypeScriptPythonGoRESTGraphQLAPI designRelational databasesNoSQLAWSGCPAzureDockerKubernetesCloud-native architectureCI/CDDevOpsObservabilitySecurityModel servingData pipelinesTechnical leadershipMentoring

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Set AI platform architecture
The role leads technical direction across frontend, backend, infrastructure and platform services. Architectural decisions must weigh scalability, performance, reliability, security, cost and time-to-market.
Build production agent infrastructure
The posting calls for infrastructure to securely run and orchestrate AI agents, plus integration of AI/ML models and data pipelines into production systems.
Deliver full-stack platform capabilities
Hands-on work includes end-to-end features, APIs, services, core libraries and developer-facing interfaces for interacting with models, agents and data.
Improve operational quality
The engineer is expected to address bottlenecks and technical risks while improving availability, fault tolerance, latency, resource use, observability and developer experience.
Own delivery and raise engineering standards
The role owns initiatives through deployment and iteration, leads design and code reviews, mentors engineers and collaborates with product, ML, data and engineering teams.

Likely challenges to think through

Containing risk in agent execution
A likely design challenge is preserving useful agent capabilities while limiting security and reliability risks in production, especially given the company's stated high-stakes use cases. Prepare to explain isolation, access controls, failure handling and observability as design considerations, without assuming a specific existing architecture.
Balancing competing platform goals
Scalability, latency, availability, resource use, cost and delivery speed may pull architectural choices in different directions. Prepare a concrete example of how to compare options, make a decision with incomplete information and revisit it using operational evidence.
Keeping integrations adaptable
The posting asks for modular components that can evolve with AI technology while integrating models and data pipelines. A plausible difficulty is avoiding tight coupling without overengineering; prepare to discuss interface boundaries, testing and a staged rollout approach.
Staying hands-on while leading across teams
The breadth of full-stack delivery, technical direction, reviews and mentorship may create competing demands on a principal engineer's time. Prepare an example showing how to retain ownership of critical implementation decisions while enabling other engineers and aligning product, ML and data stakeholders.

Practice your answers

Technical practice

How would you architect a production platform for complex AI agent workloads?

What to coverThe role calls for secure, scalable, reliable, and maintainable systems across platform services. Preparation suggestion: explain component boundaries and how you would balance performance, fault tolerance, cost, and delivery time.

How would you orchestrate agents when a step fails or must be retried?

What to coverThe posting calls for infrastructure to run and orchestrate agents in production. Preparation suggestion: discuss execution state, timeouts, retries, and how you would prevent duplicate or unsafe actions.

How would you secure the execution of agents that can act on sensitive data?

What to coverSecure agent execution is a stated duty, and secure environments for untrusted workloads are listed as a preference. Preparation suggestion: address isolation, permissions, data access, and auditability without assuming a particular implementation at Accrete.

How would you preserve trustworthy context for agents using data from multiple silos?

What to coverAccrete describes its Knowledge Engine as a dynamic context graph that resolves data across silos into an auditable ground truth. Preparation suggestion: discuss provenance, conflicting records, updates, and how an agent could distinguish supported context from uncertain information.

How would you design a developer-facing feature for interacting with models, agents, and data?

What to coverThe role spans frontend and backend development, including developer-facing interfaces and scalable APIs. Preparation suggestion: describe a React or TypeScript interface and service boundary, then explain when REST or GraphQL would better serve the interaction.

How would you integrate a model and data pipeline into a production service?

What to coverThe posting calls for integrating AI/ML models and data pipelines and mentions familiarity with model-serving infrastructure. Preparation suggestion: cover interface contracts, version changes, testing, and the trade-offs between serving latency and resource use.

How would you choose data storage for platform and agent workloads?

What to coverThe role seeks experience with relational or NoSQL databases and describes a context graph at the center of Accrete's platform. Preparation suggestion: compare consistency, query patterns, scale, and audit needs rather than assuming the graph's storage technology.

How would you deploy and update an AI platform service safely?

What to coverThe posting names cloud platforms, Docker, Kubernetes, CI/CD, and production operations. Preparation suggestion: describe deployment checks, rollback, and how you would verify that a change remains reliable under load.

How would you diagnose rising latency and resource use in production?

What to coverImproving latency, resource utilization, reliability, and observability is an explicit responsibility. Preparation suggestion: explain what you would measure across services and model workloads, how you would isolate a bottleneck, and how you would validate a fix.

How would you evaluate an emerging AI capability for the platform roadmap?

What to coverThe role includes assessing emerging AI/ML technologies and building modular components that can evolve. Preparation suggestion: define evaluation criteria for usefulness, security, operational cost, reliability, and integration effort before recommending adoption.

Behavioural practice

Tell me about a complex initiative you owned from architecture through production.

What to coverThe posting emphasizes end-to-end ownership and accountability for technical outcomes. Preparation suggestion: use situation, action, and result to show the decisions you made, how you handled deployment, and what changed afterward.

Describe a time you made an architectural decision with incomplete information.

What to coverThe role calls for independent judgment on ambiguous problems involving scale, reliability, cost, and time-to-market. Preparation suggestion: use situation, action, and result to explain the options, the evidence available, the trade-off you chose, and how you revisited it.

Give an example of raising engineering standards through a review or mentorship.

What to coverLeading design and code reviews and mentoring engineers are stated responsibilities. Preparation suggestion: use situation, action, and result to describe the guidance you gave, how you handled differing views, and the observable effect.

Describe a time you aligned product, ML, data, and engineering stakeholders.

What to coverThe role requires cross-functional collaboration to translate business requirements into technical solutions. Preparation suggestion: use situation, action, and result to show how you clarified competing needs, communicated a decision, and measured its outcome.

Tell me about feedback that changed your technical approach.

What to coverAccrete's stated values include transparency, constructive feedback, and continuous learning. Preparation suggestion: use situation, action, and result to explain the feedback, what you changed, and what you learned without presenting the example as an employer interview expectation.

Sharpen your knowledge

Topics to prepare

Distributed platform trade-offs
Posting: The role requires system design and distributed systems expertise, balancing scalability, reliability, latency, cost and delivery time. Revision exercise: Compare two architectures for a production agent workload; explain bottlenecks, failure modes and why you would choose one.
Production agent architecture
Posting: The platform must securely run and orchestrate agents and integrate AI/ML models and data pipelines. Revision exercise: Sketch an agent request lifecycle, identifying execution boundaries, model and data interactions, and recovery from failure. This is a preparation scenario, not a specified Accrete design.
API and data design
Posting: The role calls for REST, GraphQL, service architectures, and relational and/or NoSQL databases. Revision exercise: Compare REST and GraphQL contracts for a developer-facing agent interface, then justify a data-store choice and its consistency trade-offs.
Cloud-native operations
Posting: AWS, GCP or Azure; Docker; Kubernetes; CI/CD; and production observability are listed. Revision exercise: Walk through how a service moves from a container build to deployment, monitoring and rollback, identifying operational risks at each stage.
Principal-level technical judgment
Posting: The engineer sets technical direction, leads reviews, mentors engineers and owns initiatives through production. Revision exercise: Prepare a design-review discussion that states assumptions, weighs alternatives, records a decision and explains how you would align product, ML, data and engineering teams.

Skills to revise

Build an agent-facing backend API
Posting: The role builds APIs, services and libraries using Python, Go or similar languages. Suggested exercise: Implement a small REST endpoint for submitting and checking a simulated agent task, with input validation, tests and explicit failure responses; explain how its contract could differ in GraphQL.
Build a developer-facing interface
Posting: Strong React, TypeScript or similar frontend experience is requested, alongside interfaces for interacting with models, agents and data. Suggested exercise: Build a small typed interface that submits the simulated task and displays its status, result and errors.
Containerize and deploy a service
Posting: Docker, Kubernetes, cloud platforms and CI/CD are explicit technical areas. Suggested exercise: Containerize the sample API, write a basic Kubernetes deployment and describe a CI/CD check and rollback path. No particular cloud provider or pipeline tool is specified.
Instrument and test failure handling
Posting: The role improves availability, latency, resource use and observability, and establishes testing and security practices. Suggested exercise: Add service metrics or logs and tests for a timeout or failed agent task; describe how you would detect, diagnose and limit the impact of that failure.
Lead a code-review exercise
Posting: Leading code reviews and mentoring engineers are responsibilities. Suggested exercise: Review the sample API for maintainability, security and test coverage; write prioritized feedback that explains the trade-offs rather than only prescribing changes.

Your preparation checklist

Map the stated requirements to evidence
Prepare a one-page reference to relevant education, software engineering experience, production platforms, technical leadership and AI/ML work. The posting lists 10+ years and both a “Ph.D Required” heading and a PhD-or-Tier-I-master’s qualification.
Build a principal-level project walkthrough
Select a real initiative you led from architecture through deployment and iteration. Prepare a diagram, the decisions you owned, trade-offs, cross-team contributions and measured production outcomes where available.
Practice an agent-platform design exercise
Sketch a hypothetical service that runs agents against auditable organizational data, reflecting Accrete’s stated Knowledge Engine context. Explain orchestration, data access, isolation, failure handling and trade-offs among reliability, latency, cost and delivery time; do not present the sketch as Accrete’s architecture.
Show the end-to-end developer experience
Prepare a concrete example or design spanning a developer-facing interface, API, backend service, database and model or data-pipeline integration. Be ready to explain relevant choices among the frontend, backend and cloud technologies named in the posting.
Prepare a production-readiness review
Create a short checklist for an AI workload covering security, tests, deployment, observability, availability and resource use. Attach an example of a bottleneck or operational risk you identified and how you evaluated the result, if you have one.
Rehearse technical leadership decisions
Prepare concise examples of a design or code review, mentoring, and resolving an ambiguous requirement with product, ML, data or engineering partners. State your role and how the decision affected delivery or engineering standards.
Write focused questions for Accrete
Ask about the AI Platform’s current architectural risks, boundaries between platform and ML teams, and how success is measured. Clarify the posting’s differing education statements rather than assuming which applies.

Availability last checked: 5 Nov 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in India. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…