Principal AI Architect

Techshell Software; Hyderabad; On-site; INR 4,000,000 - 7,000,000; Full time

Job summary

Techshell Software in Hyderabad is seeking a hands-on AI Engineer/Architect to own the RCG implementation track within Pathfinder. You will design federated knowledge graph patterns and lead technical discussions with senior client stakeholders while shipping production systems.

You will write production-grade code in Python, Cypher/Gremlin, and orchestration scripts, leveraging LangGraph or similar frameworks. 8–12 years’ experience in AI/ML and enterprise graph systems is expected.

Qualifications

• 8–12 years of software engineering experience with substantial AI/ML systems work.
• 100% hands-on coding, not just oversight.
• Production knowledge graphs: designed schemas, non-trivial Cypher/Gremlin/SPARQL, operated Neo4j/Amazon Neptune/TigerGraph at scale.
• Strong Python and AI tooling experience: LangGraph, LangChain or equivalent.
• Experience with KG-augmented AI, ontology-driven reasoning, or KG-augmented LLM retrieval.
• Daily use of AI coding tools (Claude Code, Cursor, Copilot, etc.) in production workflows.
• Enterprise software shipping: data governance, lineage, multi-tenancy, observability.
• CS/Math/Computational Linguistics degree from IIT/IISc/IIIT-H/BITS/NIT or international equivalent.

Responsibilities

• Design and implement RCG schemas, ontologies, and reference patterns for client engagements.
• Write production-grade Python, Cypher/Gremlin, and orchestration code.
• Use AI coding tools as a force multiplier to ship faster without sacrificing quality.
• Build and operate knowledge graphs on enterprise-scale Neo4j/Amazon Neptune.
• Design federated KG architectures for multi-tenant deployments.
• Lead technical conversations with client architects, R&D directors, and CIOs.
• Mentor senior engineers and elevate technical depth of the team.
• Contribute to patent filings, white papers, and Pathfinder roadmap.

Skills

Hands-on coding

Python

Cypher/Gremlin

Knowledge graphs

LangGraph/LangChain

AI tooling

Neo4j

Amazon Neptune

GraphRAG

KG-augmented AI

Education

CS degree from IIT/IISc/IIIT-H/BITS/NIT

Tools

Neo4j

Amazon Neptune

TigerGraph

LangGraph

LangChain

Claude Code

Cursor

GitHub Copilot

Job description

Role: AI Engineer/Architect

Exp: 8 + years

Location: Hyderabad (WFO)

About Client:

It is a US-headquartered AI consulting and data engineering firm working with Fortune 500 clients in manufacturing, logistics, CPG, and automotive. We are an AWS Advanced Consulting Partner with deep SAP, Microsoft Fabric, and industrial intelligence expertise.

Pathfinder is our proprietary Agentic Accelerator Platform. At its core is RCG (Reflexive Compound Graph), our patent-pending knowledge graph architecture that powers reasoning, governance, and multi-agent orchestration in enterprise systems. We are looking for an architect who will own RCG's evolution.

The Role

This is a hands-on architect role 100% hands-on. You will write code, design schemas, and ship production systems alongside the team. This is not a slide-deck or PowerPoint architect role. If you have not written code in the last 12 months, this role is not for you.

You will own the RCG implementation track within Pathfinder, design the federated knowledge graph patterns we deploy at clients, and lead technical conversations with senior client stakeholders. You will report directly to the Founder & CEO.

What You'll Do

• Design and implement RCG schemas, ontologies, and reference patterns for client engagements
• Write production-grade Python, Cypher/Gremlin, and orchestration code (LangGraph or equivalent)
• Use AI coding tools (Claude Code, Cursor, or equivalent) as a force multiplier to ship faster without sacrificing quality
• Build and operate knowledge graphs on Neo4j and Amazon Neptune at enterprise scale
• Design federated KG architectures (Schema KG vs. Instance KG separation) for multi-tenant deployments
• Lead technical conversations with client architects, R&D directors, and CIOs
• Mentor senior engineers and elevate the team's technical depth
• Contribute to patent filings, technical white papers, and Pathfinder roadmap

Must-Haves

• 8 to12 years of software engineering experience, with at least 34 years in AI/ML systems
• 100% hands-on actively writing code today, not directing others to write it
• Production knowledge graph experience: designed schemas, written non-trivial Cypher/Gremlin/SPARQL, operated Neo4j, Amazon Neptune, or TigerGraph at meaningful scale
• Strong Python and modern AI tooling: LangGraph, LangChain, or equivalent agent orchestration frameworks
• Hands-on experience with KG-augmented AI: GraphRAG, ontology-driven reasoning, or KG-augmented LLM retrieval
• Daily, fluent use of AI coding tools Claude Code, Cursor, Kiro, GitHub Copilot, or equivalent. You should be able to demonstrate how you use these tools to ship production code, including prompting strategy, multi-file refactoring, agentic workflows, and where you trust them vs. override them
• Enterprise software shipping experience: data governance, lineage, multi-tenancy, observability
• Computer Science, Mathematics, or Computational Linguistics degree from IIT, IISc, IIIT-H, BITS Pilani, NIT, or international equivalent

Strong Preferences

• MS or PhD in graph theory, knowledge representation, computational linguistics, or applied ML on structured data
• Open-source contributions, papers, or public technical writing
• Exposure to SAP, ERP, or enterprise data models
• Product company background: Microsoft Research India, Google, Adobe Research, Flipkart/Myntra ML, Sarvam, Krutrim, or comparable AI-first company
• AWS architecture-level depth: Bedrock, SageMaker, Neptune, Lambda, IAM

This Role Is Not For You If

• You have moved out of hands-on coding into pure architecture or management
• Your knowledge graph experience is limited to tutorials, courses, or a single side project
• You list 30+ tools on your resume with shallow depth across all of them
• You have not adopted AI coding tools (Claude Code, Cursor, Copilot, etc.) into your daily workflow, or you dismiss them as hype
• You are a pure researcher who has never shipped production systems
• You are a senior architect from a large consulting/SI firm whose deliverables are primarily slide decks