AI Systems Engineer (Claude-Centered Research Systems)

Please login or register as jobseeker to apply for this job.

TYPE OF WORK

Any

SALARY

Starts at $18 per hour

HOURS PER WEEK

TBD

DATE UPDATED

Mar 12, 2026

JOB OVERVIEW

AI Systems Engineer (Claude-Centered Research Systems)

Overview

Upside Auction is building a structured AI infrastructure to power real estate auction intelligence at scale as we expand into more states and properties.

We are seeking our first dedicated AI Systems Engineer / AI Engineer to architect and implement a Claude-centered research system that automates:

1. Foreclosure and lien research
2. Title and legal document analysis
3. Court notice and deed parsing
4. Property-level research and information gathering
5. CRM and workflow automation tied to AVP and HubSpot

We are looking for a builder who can design, deploy, and operationalize real-world LLM systems inside a production workflow, with hands-on implementation, setup, and virtual environment deployment experience.

Anthropic Claude will serve as the central orchestration layer (“bot manager”) for research and decision-support processes, coordinating AI agents to perform tasks.

What You’ll Build

You will design and deploy a scalable AI research infrastructure that:

1. Acts as an internal AI research assistant for property and auction data
2. Automates foreclosure, lien, and broader legal/title research workflows
3. Coordinates AI agents to:
3.1 Retrieve documents from multiple external sites and tools
3.2 Summarize and validate results
3.3 Flag mismatches (e.g., wrong document numbers, wrong property)
4. Parses and extracts structured data from:
4.1 Court filings and notices
4.2 Notices of sale
4.3 Deeds and title documents
4.4 Legal PDFs and scanned documents
5. Integrates directly with:
5.1 Internal data sources and research tools
5.2 AVP workflows
5.3 HubSpot CRM
6. Maintains structured logs, audit trails, and validation checks so humans can quickly review and correct edge cases
7. Optimizes cost and performance of LLM usage as we scale to hundreds of properties

You will help move Upside Auction from manual, research-heavy workflows (multiple people, many minutes per property) to structured, AI-assisted research systems.

Key Responsibilities

1. Architect Claude-centered LLM workflows for legal/property research and document handling
2. Design system prompts and multi-step reasoning chains for:
2.1 Document retrieval
2.2 Structured extraction
2.3 Validation and error detection
3. Build document parsing pipelines (PDF, OCR, text extraction, structured output)
4. Implement:
4.1 Error handling and retry logic
4.2 Validation layers to catch issues like mis-labeled or mis-indexed documents
4.3 Task feedback loops between AI agents and the Claude “manager” bot
5. Deploy and maintain services in virtual/cloud environments (virtual terminals/VMs)
6. Integrate LLM outputs into:
6.1 HubSpot CRM workflows
6.2 AVP and internal automation systems
6.3 Internal folders and knowledge repositories
7. Implement monitoring, logging, and cost tracking for AI usage
8. Document:
8.1 System architecture
8.2 Installation and configuration steps (virtual terminals/environments)
8.3 Operational playbooks for non-technical tea ---------- mbers
9. Work directly with leadership and research staff to translate current manual research processes into robust AI-powered workflows

You will own this function end-to-end and will be building this as a fresh project.

Required Skills & Experience

1. Production experience implementing Anthropic Claude (real workflows, not demos)
2. Experience with OpenAI GPT/Codex or comparable LLM APIs
3. Proven ability to design and ship production systems independently
4. Strong skills in Python (preferred) or JavaScript/TypeScript
5. Comfortable with Linux, virtual environments, and cloud/VM deployment
6. Experience building API integrations and document processing pipelines (PDF/OCR/structured extraction)
7. Working knowledge of databases and structured data modeling
8. Ability to translate manual business processes into automated, agent-driven workflows
9. Clear technical documentation and communication skills

Nice-to-Have

1. Experience designing agentic or multi-step LLM systems
2. Familiarity with RAG, vector databases, or retrieval systems
3. Experience integrating with HubSpot or similar CRMs
4. Exposure to LLMOps / monitoring / prompt versioning
5. Background in real estate, foreclosure, title, or legal-tech workflows
6. Startup experience with full ownership of systems

Send us your resume and portfolio to apply.

VIEW OTHER JOB POSTS FROM:
SHARE THIS POST
facebook linkedin
  BENCHMARKS  
Loading Time: Base Classes  0.0011
Controller Execution Time ( Jobseekers / Job )  0.0357
Total Execution Time  0.0375
  GET DATA  
No GET data exists
  MEMORY USAGE  
1,527,664 bytes
  POST DATA  
No POST data exists
  URI STRING  
jobseekers/job/AI-Systems-Engineer-Claude-Centered-Research-Systems-1583126
  CLASS/METHOD  
jobseekers/job
  DATABASE:  onlinejobs (Jobseekers:$db)   QUERIES: 13 (0.0281 seconds)  (Hide)
0.0003   SELECT *
                                
FROM exrates
                                WHERE rate_name 
'USD-PHP' 
0.0016   SELECT *
FROM `employer_jobs`
WHERE `job_id` = 1583126
 LIMIT 1 
0.0009   SELECT *
FROM `employers`
WHERE `employer_id` = 753552
 LIMIT 1 
0.0009   SELECT COUNT(*) AS `numrows`
FROM `t_thread` `t`
LEFT JOIN `t_thread_misc` `miscON `t`.`id` = `misc`.`thread_id`
WHERE `t`.`job_id` = 1583126
AND `misc`.`idIS NULL 
0.0005   SELECT e.business_namee.logoe.websitee.rebill_datee.date_added member_datehitsDATEDIFF('2026-04-16',ej.date_added) duration_daysDATEDIFF('2026-04-16',e.rebill_date) duration_rebillej.*, e.deactivate FROM employers eemployer_jobs ej WHERE e.employer_id ej.employer_id AND
                                   ((
e.user_level >= '500' AND ej.date_added <= e.rebill_date)
                                   OR 
e.employer_id '' OR (ej.date_approved <> '2000-01-01' and DATEDIFF('2026-04-16',ej.date_added) <= 14 ))
                                   AND 
e.deactivate != AND ej.deleted AND job_id '1583126' 
0.0008   SELECT *
FROM `employer_jobs_skills` `ejs`
LEFT JOIN `skills_categories` `scON `ejs`.`skill_id` = `sc`.`id`
WHERE `job_id` = 1583126 
0.0011   UPDATE employer_jobs SET hit_counts '***Feb-19-2026=131***Feb-20-2026=33***Feb-21-2026=14***Feb-22-2026=6***Feb-23-2026=16***Feb-24-2026=1***Feb-25-2026=11***Feb-26-2026=2***Feb-27-2026=5***Feb-28-2026=8***Mar-01-2026=7***Mar-02-2026=9***Mar-03-2026=6***Mar-04-2026=4***Mar-05-2026=3***Mar-06-2026=1***Mar-07-2026=1***Mar-08-2026=2***Mar-10-2026=1***Mar-11-2026=1***Mar-12-2026=25***Mar-13-2026=92***Mar-14-2026=17***Mar-15-2026=11***Mar-16-2026=29***Mar-17-2026=7***Mar-18-2026=9***Mar-19-2026=9***Mar-20-2026=3***Mar-21-2026=2***Mar-22-2026=9***Mar-23-2026=7***Mar-24-2026=4***Mar-25-2026=6***Mar-26-2026=7***Mar-27-2026=3***Mar-29-2026=2***Mar-30-2026=1***Mar-31-2026=1***Apr-01-2026=4***Apr-02-2026=2***Apr-03-2026=3***Apr-05-2026=4***Apr-06-2026=4***Apr-07-2026=2***Apr-09-2026=1***Apr-10-2026=1***Apr-11-2026=1***Apr-16-2026=1' WHERE job_id'1583126'  
0.0006   UPDATE employer_jobs SET monthly_hits '***Feb-2026=227***Mar-2026=279***Apr-2026=23' WHERE job_id'1583126'  
0.0013   SELECT date_sent FROM jobseeker_sent_emails WHERE jobseeker_id '' AND job_id '1583126' AND status LIKE 'sent%' ORDER BY id DESC  
0.0003   SELECT *
FROM `employer_jobs_skills` `ejs`
LEFT JOIN `skills_categories` `scON `ejs`.`skill_id` = `sc`.`id`
WHERE `job_id` = 1583126 
0.0191   SELECT COUNT(*) AS `numrows`
FROM `employer_jobs`
WHERE `employer_id` = '753552'
AND `date_added` >= '2022-06-08' 
0.0004   select from teasers 
0.0003   SELECT FROM skill_categories WHERE skill_cat_id='' 
  HTTP HEADERS  (Show)
  SESSION DATA  (Show)
  CONFIG VARIABLES  (Show)