Principal Architect - AI & Full Stack | 10-15 Years | Bengaluru

About the role

You will own the architecture and delivery of the core autonomous systems that power Neuron7's Resolution Intelligence platform. You will translate ambiguous service problems into scalable, production-grade agentic workflows that directly impact enterprise service metrics. You will define and enforce the technical vision for multi-agent orchestration, memory, and data intelligence across the product. You will mentor engineers and elevate the craftsmanship of backend and AI systems within the team. You will ensure that the platform balances rapid experimentation with the reliability required by mission-critical service environments. You will collaborate closely with product and domain experts to align technical decisions with high-value business outcomes. You will champion best practices in deterministic design, evaluation, and continuous adaptation of AI-driven services.

Key facts

What you'll do

Architect and operate scalable multi-agent systems using planner-executor patterns and deterministic flow control to meet strict enterprise SLAs.

Build and govern a hierarchical memory stack (Ephemeral, Semantic, Structured) that allows Neuro to maintain context and learn from outcomes across thousands of product lines.

Model the Knowledge Graph infrastructure to unlock tribal knowledge and enable GraphRAG pipelines that reduce hallucinations to zero in mission-critical environments.

Create the Agentic Platform as a product, including tool registries, identity-resolved profile services, and standard A2A protocols to accelerate agent deployment.

Own production readiness end-to-end by implementing intelligent model routing, semantic caching, and streaming to consistently meet p95 latency targets under two seconds.

Establish AgentOps architectural standards for continuous evaluation and trajectory analysis, embedding circuit breakers and human-in-the-loop escalation for high-stakes decisions.

Lead PEFT strategy using LoRA-based fine-tuning to adapt open-source LLMs to niche service terminologies and deploy them efficiently via multi-adapter serving.

Define and evolve the full LLMops/Agentops lifecycle, including distributed tracing with OpenTelemetry, cost-per-goal optimization, and resilience patterns.

Translate vague diagnostic requirements into structured resolution workflows, navigating ambiguity to turn unclear problems into well-defined system improvements.

Balance innovation with platform consistency, guiding teams from ad-hoc scripts to reusable modules that form a Golden Path for developers.

Identify problems that meaningfully impact EBITDA by aligning technical capabilities with enterprise service economics and strategic priorities.

Mentor engineers on failure modes, resilience, and the disciplined engineering required for autonomous systems in complex service domains.

Champion data and evaluation rigor to ensure that agent behavior is measurable, explainable, and continuously improvable.

Partner with service domain experts to encode expert insights into scalable AI workflows that enhance first-call resolution, turnaround time, and service margins.

Drive architectural decisions that enable rapid yet responsible scaling of AI capabilities across high-tech, manufacturing, and medical device service environments.

Requirements

Master's or PhD in Computer Science or a related technical discipline.

Minimum of 10-12 years of industry experience with at least 3-5 years delivering production-grade AI, ML, or Generative AI solutions at scale.

Demonstrated ability to architect multi-agent orchestration systems using frameworks such as LangGraph, AutoGen, or PydanticAI.

Proven experience building and managing complex, hierarchical memory stacks that support long-running context and learning.

High proficiency in Python and at least one compiled language such as Java, Go, Rust, or C++ for performance-critical components.

Expertise in Knowledge Graph modeling using technologies like Neo4j or Memgraph and in constructing GraphRAG pipelines that outperform naive RAG.

Strong operational discipline across the LLMops and Agentops lifecycle, including distributed tracing with OpenTelemetry, trajectory evaluation, and cost optimization.

Hands-on experience with parameter-efficient fine-tuning methods such as LoRA and QLoRA, and with multi-adapter serving frameworks.

Clear understanding of platform engineering principles to build Golden Path abstractions that accelerate developer velocity.

Ability to clarify problems before proposing solutions, weighing trade-offs between structured and unstructured knowledge retrieval.

Comfort with navigating poorly defined problems and turning ambiguous requirements into concrete system goals.

Commitment to building for failure with robust guardrails, human-in-the-loop pathways, and continuous evaluation mechanisms.

Focus on measurable outcomes, including latency, cost-per-goal, first-call resolution, turnaround time, and service margins in enterprise contexts.

Nice to have

Experience contributing to open-source agentic frameworks or data intelligence tooling.

Background in high-tech, manufacturing, or medical device service domains and familiarity with service intelligence KPIs.

Practical notes

This is a full-time role based in Bengaluru. The position requires significant collaboration across engineering, product, and customer-facing teams and may involve travel within client environments where applicable. Candidates must be eligible to work in India without sponsorship.