AI Engineer

Aurigo Software Technologies

39d

About The Position

This role involves two primary responsibilities: field deployment and model development. As a builder on the deployment team, you will configure and integrate Aurigo AI agents within customer environments, connecting the platform with existing data systems and resolving technical challenges in government IT settings. You will collaborate directly with agency data teams, IT staff, and system administrators. The second part of the role is model development, where you will build and fine-tune ML models for specific capital program data when pre-built AI is insufficient. This includes designing data pipelines, engineering features, running training cycles, evaluating performance, and contributing reusable models back to the Aurigo platform.

Requirements

3+ years building and deploying ML models in production.
Proficiency in Python ML stack: scikit-learn, PyTorch, TensorFlow, or HuggingFace Transformers.
Experience with NLP techniques applied to document-heavy data: text classification, named entity recognition, embedding models, semantic search.
Working knowledge of LLM fine-tuning, RAG architecture, or prompt optimization in domain-specific applications.
Hands-on experience building data pipelines for unstructured or semi-structured data (PDFs, XML exports, structured logs) and transforming them into model-ready features.
REST API integrations and comfort with the engineering work of connecting enterprise systems.
Ability to work independently in ambiguous field environments; diagnose and build without waiting for a perfectly scoped ticket.

Nice To Haves

Familiarity with MLOps practices: model versioning, evaluation pipelines, monitoring for drift, and retraining workflows in production.
Experience with construction, infrastructure, or capital program data (cost codes, schedule structures, contract document formats, or similar domain data).
Prior work in a field deployment, systems integration, or technical consulting role.
Familiarity with vector databases (Pinecone, Weaviate, pgvector) or knowledge graph approaches for domain-specific retrieval.
Experience in government or regulated environments (navigating IT procurement, access controls, and security requirements).
Public Trust clearance eligibility.

Responsibilities

Configure and deploy Aurigo AI agents within customer Masterworks environments, tailoring agent behavior, workflows, and outputs to each agency's specific requirements.
Build and maintain data integrations between Masterworks and agency systems, including scheduling tools, cost systems, financial management platforms, document management, GIS, and agency data warehouses.
Develop scripts and lightweight automation to streamline agency data workflows, reduce manual handoffs, and prepare data for agent consumption.
Work with agency IT staff, data stewards, and system administrators to navigate access, permissions, and integration constraints in government technology environments.
Troubleshoot deployment issues in the field, diagnosing root causes, implementing fixes, and documenting solutions for reuse across future deployments.
Design and train custom ML models on capital program data, such as cost overrun prediction, schedule risk scoring, anomaly detection in project financials, and document classification, deployed as intelligence layers inside Aurigo agents.
Build feature engineering pipelines from Masterworks and connected systems, transforming raw program data into structured, model-ready inputs.
Fine-tune or adapt large language models for infrastructure-specific tasks like RFI response drafting, submittal compliance review, meeting minute summarization, and specification and contract parsing.
Build data preprocessing pipelines for unstructured construction documents (PDFs, field reports, RFI logs, change order packages), transforming them into structured, model-ready datasets.
Develop and maintain model evaluation frameworks; monitor production model performance, identify drift, retrain as needed, and document performance metrics for each deployment.
Contribute models, pipelines, and reusable components back to the Aurigo product team, building the platform's AI capability from field learnings.