Agentic AI Architect (Only local to TX)- C2C Requirements

Tech Mirrors; Austin (TX); Hybrid; USD 80,000 - 113,000; Part time

Job summary

Tech Mirrors in Austin, TX is seeking a Principal Architect with deep AI and agentic systems to design and deliver production-grade, server-hosted AI solutions. This role blends architecture leadership with hands-on engineering, focusing on LLM orchestration, memory management, RAG, and enterprise-grade agent workflows.

The ideal candidate has proven experience building scalable agentic systems in enterprise environments, not just experiments, and can design systems that integrate business

Qualifications

• Agentic AI / LLM systems: server-based agentic systems experience.
• Memory and context management: design memory for sessions and tasks.
• Advanced Retrieval / RAG: vector/graph DBs, embeddings, retrieval design.
• Code intelligence/technical analysis: ingest code into AI workflows; AST parsing.
• Programming depth: strong Python/Java; familiarity with AWS/AI platforms.
• Frameworks/platforms: LangGraph, LangChain, Copilot SDK, Bedrock.

Skills

Agentic AI/LLM Systems

Memory/Context Management

Advanced Retrieval/RAG

Code Intelligence/Technical Analysis

Programming Depth (Python/Java)

Frameworks/Platforms (LangGraph/LangCH

Tools

Vector Databases

Graph Databases

Embeddings/Retrieval Design

AST Parsing (Tree-sitter)

Redis / Apache Ignite

SOLR / Lucene

Bedrock / Cloud AI

Copilot SDK

Job description

Austin, TX (Hybrid – 3 days/week)

6 months (Possible extension)

Rate: $70/hr c2c (Max)

Note: Please ensure that candidates complete the home exercise. Kindly communicate this requirement to the candidates in advance.

Principal Architect with Deep AI & Agentic Systems – Detailed JD

Principal Architect with deep AI and agentic systems expertise to design and deliver production-grade, server-hosted AI solutions. This role requires a balance of architecture leadership and hands‑on engineering, with a strong focus on LLM orchestration, memory management, advanced retrieval (RAG), and enterprise-grade agent workflows.

The ideal candidate has proven experience building scalable agentic systems in enterprise environments, not just experimental or IDE-based solutions, and can design systems that integrate business context, code intelligence, and human-in-the-loop workflows.

Must-Have Skills

Recruiters should screen strongly for the following:

1) Agentic AI / LLM Systems

• Proven experience building agentic systems that run-in server environments
• Experience with multi-step orchestration, tool calling, and workflow design
• Hands‑on implementation experience — not just conceptual AI knowledge

2) Memory and Context Management

• Experience designing short-term and long-term memory for AI systems
• Understanding of how to manage context across sessions, tasks, and workflows
• Familiarity with state handling, summarization, persistence, and context optimization

3) Advanced Retrieval / RAG

Strong experience with advanced RAG architectures

Hands‑on work with:

• Vector databases
• Graph databases
• Embeddings and retrieval design
• Experience with hybrid retrieval patterns and building richer context for AI systems

4) Code Intelligence / Technical Analysis

• Experience ingesting source code and documents into AI/retrieval workflows
• Ability to support: impact analysis, dependency mapping, technical analysis, code relationship/call graph understanding
• Experience with AST parsing using tools such as Tree-sitter or similar

5) Programming / Engineering Depth

Strong hands‑on experience in:

• Python or Java
• Agentic Platforms, RAG/GraphDB, LLMs
• Caching – near and distributed (Redis ,Apache Ignite etc)
• API Token management optimization techniques is a plus
• Experience with SOLR, Lucene or equivalent is a plus
• AWS Cloud Experience is a plus

Enough engineering depth to contribute meaningfully to implementation and technical design

6) Frameworks / Platforms

Experience with one or more of the following is important:

• LangGraph
• LangChain
• GitHub Copilot SDK or similar
• CLI-based/server-hosted execution models
• Managed cloud AI platforms such as Bedrock or similar offerings