Agentic AI Developer

Machine Learning Technologies, LLC; Atlanta (GA); On-site; USD 120,000 - 160,000; Full time

Job summary

Machine Learning Technologies, LLC in Atlanta seeks an experienced AI/ML engineer to drive embedded discovery and solution delivery. You will translate business requirements into technical specifications, acceptance criteria, and incremental delivery plans, while identifying when deterministic logic, automation, or AI agents are appropriate.

Responsibilities include building production-grade backend services, Databricks notebooks, and SQL transformations, with a focus on data quality,

Qualifications

• Proficiency in translating business needs into technical specs and acceptance criteria.
• Experience building production-grade backend services and data-processing components.
• Familiarity with AI governance, tool contracts, and safety considerations.

Responsibilities

• Partner with business users to understand workflows, pain points, and outcomes.
• Translate requirements into user stories, testable criteria, and delivery plans.
• Develop bounded AI capabilities with structured outputs, prompt versioning, and evidence-backed results.
• Build production-quality backends, Databricks notebooks/jobs, and SQL transformations for data processing.
• Implement data-quality checks, schema validation, deduplication, and traceable processing rules.
• Collaborate on reviewer-facing interfaces and end-to-end application behavior.
• Ensure least-privilege access, secure authentication, and secrets management.

Skills

AI/ML concepts

Backend development

APIs

Data processing

Testing/QA

Education

Bachelor's degree in CS/EE/Math

Tools

Databricks

SQL

Python

APIs

Job description

Job Description:

• Embedded Discovery and Solution Delivery: Partner directly with business users and subject-matter experts to understand workflows, pain points, exceptions, and desired outcomes.
• Translate business requirements into technical specifications, user stories, testable acceptance criteria, and incremental delivery plans.
• Identify when deterministic logic, conventional automation, retrieval, or an AI agent is the appropriate solution.
• Build working vertical slices that demonstrate business value rather than disconnected proofs of concept.
• Communicate technical trade-offs, risks, dependencies, and decisions clearly to technical and nontechnical stakeholders.
• Agentic AI and LLM Application Engineering: Develop bounded AI capabilities for research, explanation, recommendation, and workflow assistance. Implement structured outputs, context retrieval, tool/function calling, prompt versioning, and output validation. Design tool contracts with explicit permissions, schemas, failure behavior, and audit requirements. Ground AI responses in approved, versioned sources and preserve links to supporting evidence. Implement human review, escalation, abstention, and fallback behavior.
• Enforce authorization boundaries in application code and tool permissions—not through prompts alone. Evaluate model and framework options against approved enterprise requirements for quality, security, latency, and cost.
• Software and Data Engineering: Build production-quality backend services, APIs, data-processing components, and workflow integrations. Develop Databricks notebooks/jobs and SQL transformations for ingestion, normalization, matching, comparison, and validation.
• Implement data-quality checks, schema validation, deduplication, reconciliation, and versioned processing rules. Maintain reproducibility through pinned rule sets, input references, and traceable execution records.
• Collaborate on reviewer-facing interfaces and end-to-end application behavior.
• Engineer reliable integrations using appropriate retry, idempotency, timeout, and recovery patterns.
• Quality, Evaluation, and Verification: Create unit, integration, end-to-end, regression, and negative tests. Develop representative evaluation datasets with domain experts. Assess AI outputs for grounding, correctness, usefulness, safety, and appropriate abstention. Test unauthorized tool use, prompt injection, sensitive-data exposure, and failure scenarios. Trace delivered functionality and verification evidence to acceptance criteria. Review AI-generated code with the same rigor as manually written code.
• Security, Governance, and AI-DLC: Follow the applicable AI-DLC lifecycle, approved tooling paths, review requirements, and release controls.
• Implement least-privilege access, secure authentication, secrets management, and data-access restrictions.
• Use synthetic, mock, or approved de-identified development data where required.
• Maintain architecture decisions, prompt/tool documentation, approval evidence, and delivery records.
• Escalate unresolved data classification, privacy, model approval, and risk decisions to accountable owners.
• Preserve human authority over final business decisions and production promotion.
• Operational Readiness: Instrument application behaviour, retrieval, tool calls, model usage, failures, and dependencies.
• Establish prompt/model traceability and appropriate telemetry protections.
• Implement usage and cost attribution, monitoring, s, and agreed reliability measures.
• Produce deployment guidance, rollback procedures, runbooks, and support handover materials.
• Work with platform and operations teams while retaining delivery-team accountability for application quality and remediation.

We are an equal opportunity employer. All aspects of employment including the decision to hire, promote, discipline, or discharge, will be based on merit, competence, performance, and business needs. We do not discriminate on the basis of race, color, religion, marital status, age, national origin, ancestry, physical or mental disability, medical condition, pregnancy, genetic information, gender, sexual orientation, gender identity or expression, national origin, citizenship/ immigration status, veteran status, or any other status protected under federal, state, or local law.