Summer research internship

Additive Manufacturing Printability Intelligence

Build the data and evaluation layer for automated printability screening across drawings, 3D model exchange, metadata, and expert review—using vocabulary aligned with ISO/ASTM AM process families where it helps consistency.

Apply by EmailBack to Careers

Core mission

Auditable datasets and evaluations for manufacturing AI.

You will shape datasets, printability labels, annotation workflows, and evaluation methods behind Accio3D's printability intelligence—bridging drawings, CAD exchange models, and part metadata with clear, traceable decisions about AM feasibility.

What you will do

Research, data systems, and AM evaluation.

Data pipeline and part representation

  • Ingest engineering data: 2D drawings, PDFs, STEP/STL and other CAD exchange geometry, metadata, supplier forms, and structured tables.
  • Extract geometry and manufacturing attributes: envelope, volume, surface area, local wall thickness, cantilevered features, internal voids, slenderness, and support-relevant topology.
  • Normalize units, formats, material designations (including supplier naming vs. standard references), tolerances, and surface texture callouts.
  • Flag incomplete, ambiguous, or conflicting records; build feature tables for evaluation and training.

Labeling schema and annotation workflows

  • Define printability classes (e.g. feasible, not feasible, needs redesign, insufficient data).
  • Encode risk factors: thin sections, orientation-sensitive features, tolerance and datum sensitivity, support burden, thermal distortion, finishing load, material–process fit, and build-volume limits.
  • Write annotation guidelines; track provenance, evidence, assumptions, and confidence.
  • Measure agreement between annotators and surface ambiguous cases.

Printability evaluation and algorithms

  • Implement rule-based checks for hard constraints: envelope, minimum feature size, wall thickness, and missing critical inputs.
  • Prototype scoring or classification for manufacturability risk; compare rules, expert labels, and model-assisted outputs.
  • Build evaluation sets and metrics; analyze false positives and false negatives.

Explainability and evaluation artifacts

  • Produce reports that tie classifications to drawing, CAD, and metadata evidence.
  • Help set acceptance criteria and regression benchmarks as logic evolves.

Research synthesis and AM domain understanding

  • Review design-for-AM, manufacturability screening, CAD feature extraction, and process constraints from standards and literature.
  • Map domain knowledge into label definitions and checks.
  • Contrast constraints across ISO/ASTM additive manufacturing process categories (e.g. material extrusion, powder bed fusion, vat photopolymerization, binder jetting, directed energy deposition) and note what generalizes vs. what is process-specific.

Required qualifications

  • Pursuing or recently finished a degree in mechanical, manufacturing, materials, industrial, or computer engineering; robotics; computational design; or a related field.
  • Interest in additive manufacturing, design for AM, CAD/CAM, or manufacturing automation.
  • Comfort reasoning about engineering constraints from structured technical data.
  • Programming in Python; workable skill with tables, files, APIs, or data scripts.
  • Careful labeling and validation habits; clear documentation of assumptions and edge cases.
  • Comfort in human-in-the-loop workflows with engineers and product.

Preferred qualifications

  • 3D CAD exchange and mesh formats (e.g. STEP, STL; others as needed).
  • CAD tools or geometry libraries (SolidWorks, Fusion 360, Onshape, FreeCAD, OpenCascade, trimesh, meshio, or similar).
  • Computational geometry, mesh processing, or feature extraction.
  • AM constraint concepts: thin walls, overhangs, supports, anisotropy and build orientation, tolerances, surface texture, residual stress and distortion, post-processing.
  • ML, computer vision, LLM evaluation, weak supervision, or active learning.
  • Annotation playbooks, labeling QA, or benchmark datasets.
  • OCR, document parsing, engineering drawings, geometric dimensioning and tolerancing (per ASME Y14.5 / ISO GPS), or technical data packages.
  • pandas, NumPy, scikit-learn, SQL, Databricks, MLflow, or similar.
  • Research in AM, design automation, computational geometry, materials informatics, or manufacturing AI.

Traits that matter

Sound judgment beats credentials alone.

  • Care over speed: poor labels undermine the whole evaluation stack.
  • Judgment on true infeasibility vs. uncertainty.
  • Tolerance for messy legacy data.
  • Visible assumptions and missing data.
  • Outcome-oriented: a usable data and evaluation base, not only experiments.

Possible deliverables

Concrete internship outputs.

  • Printability schema v1: labels, risk axes, severity, confidence.
  • Curated benchmark set with sources, labels, rationale, and QA.
  • Ingestion prototype producing structured feature records.
  • Baseline evaluator with constraint checks and short rationales.
  • Evaluation summary: coverage, ambiguity, rule/model behavior, failure modes, data priorities.
  • Labeling playbook: onboarding, review, QC, escalation.

Interested in the role?

Send a résumé, a short note on fit, and examples of research, code, geometry, or data work.

Apply by Email