Ai systems engineer

Mobius By Gaian · Hyderabad g.p.o., India

About the role

We are hiring multiple AI Systems Engineers to join Mobius Research Lab in India and help build a new class of AI platform: one that goes beyond chatbots, shallow agents, and one-off automation workflows.

This role is for a deeply technical engineer or applied researcher who wants to work on the foundations of emerging AI systems: structured AI labor, knowledge representation, agent orchestration, graph-based reasoning, secure runtime execution, workflow compilation, model integration, and enterprise-grade validation. You will help build systems that can ingest complex real-world information, convert it into structured machine-understandable form, reason over it, produce executable plans, validate outcomes, and operate safely across modern cloud and AI infrastructure.

This is not an ordinary AI application role. It is a chance to work close to the platform layer where the next generation of AI systems will be defined: reliable, composable, governable, secure, and capable of operating at world scale.

Who we are looking for

We are looking for candidates who combine deep academic preparation with hands-on engineering ability. The ideal candidate has strong foundations in computer science, AI, distributed systems, data systems, graph reasoning, or secure platform engineering, and is excited by unusually complex, open-ended technical challenges.

-Education requirement: Ph D or M. Tech only, from premier or highly reputed institutions with strong computer science, AI, systems, mathematics, data science, or engineering programs.

-Experience requirement: 5 to 7 years of relevant experience in AI systems, backend platforms, distributed systems, data systems, ML infrastructure, knowledge graphs, security, or enterprise automation.

-Research depth: Ability to read dense technical material, reason from first principles, formulate abstractions, and convert research-grade ideas into working systems.

-Builder mindset: Strong ability to prototype quickly, validate rigorously, harden what matters, and take responsibility for correctness.

-Ambition: A desire to do extraordinarily complex and challenging work with the potential to make an impact on the world stage.

Candidates from institutions such as IISc, IITs, IIIT-H, ISI, CMI, top NITs, BITS Pilani, and internationally comparable universities are strongly encouraged to apply. Equivalent evidence of exceptional research and engineering depth may be considered only where the academic bar is clearly met.

What you will work on

You will work on the core platform layer that turns AI reasoning into durable, auditable, executable capability. Your work may include:

AI compiler and transformation pipelines: Build pipelines that take documents, APIs, workflows, schemas, policies, and domain knowledge, then transform them into structured internal representations that downstream AI agents and services can use.

Structured LLM labor systems: Design systems where LLMs perform accountable work: decomposition, classification, mapping, extraction, schema generation, validation, repair, synthesis, and explanation.

Knowledge ingestion and representation: Create pipelines that ingest enterprise documents, Open API specs, BPMN workflows, JSON/YAML files, standards, contracts, and operational data, then convert them into typed knowledge graphs, semantic objects, lineage records, and reusable execution context.

Agentic orchestration: Build multi-agent and tool-using systems that can plan, call tools, coordinate tasks, manage intermediate state, recover from failure, and produce auditable outputs.

Graph reasoning and validation: Develop graph validators, compatibility checkers, state-transition checks, provenance verifiers, dependency analyzers, and repair workflows to make sure AI-generated structures are internally consistent and execution-ready.

Secure AI runtime integration: Connect AI reasoning systems to execution surfaces such as Kubernetes, workflow engines, Git Ops, serverless tasks, GPU/TPU jobs, confidential VMs, policy engines, and enterprise APIs.

Evaluation and observability: Build test harnesses, evaluation suites, trace systems, model-output validators, regression checks, quality gates, and metrics that make AI behavior measurable and improvable.

The kinds of problems you will solve

You will work on hard, high-value AI engineering problems such as:

How do we turn unstructured knowledge into reliable structured objects?

How do we make LLM output deterministic enough for enterprise workflows?

How do we prevent agents from becoming loose, untraceable chains of prompts?

How do we validate AI-generated plans before they touch production systems?

How do we preserve provenance across documents, model calls, graph transformations, and runtime actions?

How do we safely connect AI agents to APIs, infrastructure, workflows, and business processes?

How do we route work across CPUs, GPUs, TPUs, and secure compute environments based on cost, priority, and risk?

How do we build AI systems that can improve themselves without becoming uncontrolled or opaque?

How do we make AI engineering feel less like prompt crafting and more like building a real operating platform?

Technical stack and expertise

We expect candidates to be comfortable with a modern AI-platform engineering stack. You do not need to know our internal architecture before joining; we care about your ability to learn quickly, reason deeply, and build with discipline.

Languages: Python is essential. Type Script is highly valuable. Go, Rust, or Java is a plus.

LLM and AI systems: Experience with frontier-model APIs, open-source models, tool calling, structured generation, function calling, RAG, embeddings, model routing, evaluation frameworks, and prompt orchestration.

Data and graph systems: Postgres, graph databases, vector databases, object storage, search systems, event logs, metadata stores, lineage systems, and knowledge graph tooling.

Schemas and contracts: JSON Schema, Open API, Async API, Protobuf, YAML, XML, BPMN, DITA, policy-as-code, contract testing, and schema validation.

Runtime and infrastructure: Kubernetes, Docker, Argo CD, Helm, Kustomize, infrastructure-as-code, CI/CD, workflow orchestration, GPU scheduling, cloud services, and observability tooling.

AI infrastructure: GPU/TPU workloads, model serving, batch inference, fine-tuning, Lo RA, vector search, model evaluation, distributed workloads, and cost-aware scheduling.

Security and governance: OIDC, RBAC, secrets management, KMS, Vault, signing, SBOMs, supply-chain security, confidential compute, secure workload execution, audit trails, and policy enforcement.

Mobius is building foundational AI platform work. This is an opportunity to help define the systems layer of emerging AI: structured model labor, graph-based knowledge, agentic orchestration, secure execution, validation, and runtime integration.

Most AI roles ask you to build features on top of models. This role asks you to help build the platform layer that makes AI useful, reliable, and valuable at scale.

If you want to work on unusually difficult AI systems, with a small high-caliber team, and with the ambition to create impact on the world stage, this is the role.