Applied AI - Technical Architect

This is a permanent, full-time role working within Insight Global's Applied AI division (IG Labs)

Location: 75% remote, 25% domestic travel

IG LABS - Program Delivery & IP Engineering

About the Role

We are looking for a Technical Architect who operates at the seam between client-facing delivery and internal IP development. You will partner directly with enterprise clients to define AI and data architecture, and then translate the hard-won patterns, decisions, and frameworks from those engagements into reusable assets that IG Labs owns and re-deploys. This is not a pure consulting role and it is not a pure product role. It is the connector between them. You will be embedded in active engagements, you will shape architecture under real constraints, and you will be accountable for making sure what we build for one client becomes leverage for the next.

Responsibilities

The Technical Baseline: Three Pillars

The Technical Architect is evaluated against the same three pillars as our FDEs, at greater depth and breadth. You are the person who sets the technical bar, so you have to clear it convincingly in all three. Interviews probe each pillar directly.

Pillar What we evaluate, and what “deep” looks like

1. System Design

For the Technical Architect, system design is broad and deep. Everything expected of an FDE (databases, data modeling, strong programming) plus the architect's canvas: cloud architecture across at least two major providers, networking and connectivity, containerization and Kubernetes/orchestration, distributed systems and scalability, reliability and observability, and security and compliance as first-class design constraints. You reason about trade-offs across the whole system, not just one workstream.

2. Traditional AI / ML

A real understanding of how models actually work, not just how to call them. Neural network fundamentals across transformers, CNNs, and RNNs — attention, tokenization, training vs. inference, loss and evaluation, overfitting and regularization. Above all, a deep working understanding of embedding spaces: how text and other modalities become vectors, what distance and similarity mean, dimensionality, and how embeddings drive retrieval, clustering, and semantic matching.

3. Applied AI / Agentic AI

Very deep, hands-on experience building production-grade agentic systems. Not demos. Orchestration and control flow, tool and function calling, RAG and context engineering, memory and state, multi-step planning, evaluation and guardrails, cost and latency management, observability, and safe deployment into real environments. You have shipped agents that real users depend on, and you know why the hard ones fail.

The difference from the FDE bar is scope: an FDE goes deep on a workstream, the Architect reasons across the whole system and across concurrent engagements.

Key Responsibilities

Client Architecture Definition

• Partner with client stakeholders (CTO, VP Engineering, Chief Data Officers) to define AI and data platform architecture aligned to their strategic objectives
• Lead architecture discovery sessions: current-state assessment, gap analysis, future-state design, and roadmap sequencing
• Produce and own architectural artifacts: reference architectures, decision records (ADRs), data flow diagrams, integration blueprints, and governance frameworks
• Define and enforce non-functional requirements across security, scalability, observability, and regulatory compliance
• Serve as the technical authority on engagements, bridging business intent and engineering execution

IG Labs Collaboration & IP Derivation

• Work hand-in-hand with the IG Labs team to identify reuse opportunities within active client engagements
• Abstract generalizable patterns from bespoke client solutions (reference architectures, design templates, evaluation frameworks, deployment playbooks) and contribute them to the IP library
• Participate in IP review cycles: peer review of contributed assets, validation of generalizability, and documentation of applicability conditions
• Actively drive toward Reuse Index targets, the primary delivery health metric tracking what proportion of each engagement is powered by pre-built IP
• Identify net-new IP opportunities from emerging client problems and partner with Labs to prototype and productize

Program Delivery

• Provide architectural oversight and technical governance across multiple concurrent client engagements
• Partner with delivery leads and project managers to ensure architectural decisions are reflected in project plans, sprint goals, and acceptance criteria
• Support pre-sales and solution design: contribute to SOWs, RFP responses, and technical proposals with defensible architectural thinking
• Mentor delivery engineers and junior architects, elevating technical craft across the team
• Flag architectural drift, scope risk, and technical debt in active engagements before they become delivery blockers

Qualifications

• 8+ years in technical architecture, solutions architecture, or principal engineering roles
• System design breadth and depth: cloud-native architecture across at least two major platforms (AWS, Azure, GCP), networking, containerization and Kubernetes, distributed systems, plus strong databases and data modeling
• Traditional AI/ML understanding: how transformers, CNNs, and RNNs work, and a deep working grasp of embedding spaces
• Hands-on experience designing and shipping production-grade agentic AI systems, not just prototypes
• Experience working directly with senior client stakeholders to define architecture
• Strong written and visual communication: architecture diagrams, decision records, executive summaries
• Demonstrated ability to abstract reusable patterns from specific implementations
• Comfortable operating with ambiguity in professional services or consulting environments
• Deep experience with RAG architectures, LLM orchestration (LangChain, LlamaIndex, or similar), and agent evaluation/observability
• Familiarity with MCP (Model Context Protocol) and its role in AI system integration
• MLOps platform design, model governance, and AI lifecycle management
• Experience with embedding/vector stores and retrieval design at enterprise scale