AI/ML Engineering Interns

About Logikality

AI-powered mortgage decision intelligence,

Lenders today lose money on every loan. Back-office costs have exploded from roughly $7,600 to over $11,000–$12,600 per loan, and 70–90% of the work is still manual despite decades of “automation.”

We’ve built a platform that:

• Processes

  180–250 pages/minute

  at ~

  98% accuracy

• Automates underwriting decisions and conditions with

  70–75% acceptance

• Validates loans against

  200+ Fannie/Freddie/investor rules

  with full explainability and auditability
• Delivers

  $1,500–2,000 savings per loan

  ,

  4–6.5 hours

  manual effort saved, and

  payback in under 30 days

  (625–1000% ROI)

15,000+ document golden test set

What you’ll work on

AI/ML Intern

• Turning a messy loan file into a

  self-organizing workspace

  : documents auto-classified, data extracted, and linked back to evidence.
• Building and evaluating models that

  compute borrower income

  across W-2s, paystubs, rental income, SSI, etc., then explain the math in auditor-ready language.
• Designing copilot-style workflows that

  generate underwriting conditions

  , track resolution, and minimize rework.
• Working on

  LLM + retrieval + rules

  pipelines that must be explainable and safe in a regulated environment.
• Using our

  15k+ document golden dataset

  plus underwriter feedback to create lender-specific playbooks and continuous learning loops.

real production data and constraints

What you’ll learn

• How to build

  decision-ready AI

  , not just models that output numbers.
• How to design for

  auditability and provenance

  as first-class citizens.
• How to ship ML in a

  highly regulated, high-stakes

  domain.
• How to work with a small, senior team where your work is

  immediately visible

  .

Who we’re looking for

traits

Must-haves

• Strong programming skills in

  Python

  (NumPy / Pandas / basic PyTorch or TensorFlow).
• Solid ML fundamentals: supervised learning, evaluation metrics, overfitting/regularization, basic NLP.
• Evidence of

  prior success

  in something hard: projects, research, competitions, open-source, startups, Olympiads, etc.
• High

  curiosity

  – you self-teach; you’ve gone down rabbit holes on your own.
• Ability to commit

  20+ hours/week

  for

  6 months

  .
• Comfortable with a

  high-intensity, unpaid internship

  where the main “compensation” is learning, network, and a serious shot at a core role.

Nice-to-haves

• Experience with

  NLP / LLMs / RAG

  , or productionizing ML systems.
• Interest/experience in

  fintech, lending, credit risk, or mortgages

  .
• Experience designing

  data/label pipelines

  and evaluation harnesses.