Project Role : Custom Software Engineer
Project Role Description : Develop custom software solutions to design, code, and enhance components across systems or applications. Use modern frameworks and agile practices to deliver scalable, high-performing solutions tailored to specific business needs.
Must have skills : AI & Data Solution Architecture
Good to have skills : NA
Minimum 15 year(s) of experience is required
Educational Qualification : 15 years full time education
Summary: Minimum 2 year experience is required in this role
Technical lead for the entire program. This person owns all architecture decisions, the AI/ML strategy, and the platform design from Day 1 through enterprise scale.
They will work alongside a small senior team and a SMEs to deliver something that does not exist anywhere in the organization today — a graph-native, agent-driven knowledge platform connecting every part of the engineering estate into a single query able intelligence layer.
Responsibilities
Own the full 9-layer architecture from connector ingestion through MCP Gateway to agent orchestration and the enterprise UI
Lead all Architecture Decision Records (ADRs) and design reviews, ensuring alignment with s enterprise constraints and zero-trust security posture
Design the LangGraph-based Core: the 13-node StateGraph, hybrid BM25+vector retrieval pipeline, context packer, and Bedrock model routing strategy
Architect the MCP Gateway — the single enforcement point for all agent tool calls — including per-agent RBAC, rate limiting, token budget enforcement, and full audit trail
Own the entity resolution architecture: blocking, matching, merge, and master index design across 9 heterogeneous source systems (GitHub, Confluence, ServiceNow, JIRA, CMDB, SonarQube, Dynatrace, REMC, RepoMapping)
Own the Bedrock model strategy: Haiku vs Sonnet tiering, cross-region inference profiles, PII/PCI guardrails, and RAGAS evaluation pipeline
Drive the tech stack approval process: present AWS services, open-source frameworks, and third-party MCP integrations to the architecture review board
Mentor AI native engineers conduct code and design reviews set standards for observability (OTel/ADOT), testing, and idempotent a pipeline patterns
Partner with the Product Owner to translate business use cases into architectural requirements across three pilot workstreams: AI Productivity, AI First Change, and AI First Run
Define Phase 3 enterprise patterns: SPIFFE/SVID zero-trust mesh, multi-tenant graph isolation, and the longitudinal learning loop
Required skills
Deep expertise in LLM orchestration frameworks — LangGraph, LangChain, or equivalent can build multi-node stateful agents from scratch
Hands-on AWS: ECS Fargate, Neptune Serverless, OpenSearch Serverless, Bedrock, ElastiCache, RDS, Secrets Manager, IAM/IRSA — not conceptual familiarity
Graph abase design: OpenCypher, Neptune, entity-relationship modeling across heterogeneous sources
Vector search and hybrid retrieval: embedding models, BM25, dense+sparse fusion, RAGAS evaluation
Python, FastAPI, and production-grade async service design
Security architecture: zero-trust networking, SPIFFE/SVID, mTLS, workload identity, PII/PCI guardrail design
Knowledge graph and entity resolution: blocking, matching, survivorship, master a management
OTel/ADOT end-to-end instrumentation and distributed tracing
Terraform or AWS CDK for infrastructure-as-code at scale
Ability to lead technical approval processes with enterprise architecture review boards
Experience presenting complex technical strategy to executive and business stakeholders
15 years full time education
