AI Researcher (Intern)

AI Researcher Intern — On-Device Model Evaluation
Type: Internship — Part-time, self-paced
Duration: 3–6 months
Location: On campus + remote
Eligibility: Current students (.edu email)
You use COTID’s on-device AI as part of your real work, and you find out where the model succeeds, where it fails, and how that difference feels from the user’s side — a bridge between what the model can do and what the product claims it does. COTID itself is your bench: you build small eval suites on one element and catalog outcomes on modest consumer hardware, because our thesis is AI that runs affordably on consumer machines, not datacenter gear.
Responsibilities

• Real-use evaluation: use the AI features on your actual coursework and build small task suites on your element — cataloging what worked, what failed, what surprised you
• Behavior analysis: study reasoning traces, tool calls, and citation behavior — when do outputs earn user trust, when do they break it
• Interaction feedback: parameter controls, streaming, suggestions, and in-editor actions during real note-taking — what helps, what distracts
• Consumer-hardware performance: measure on a modest laptop/PC — does it actually feel good at consumer spec
• Error taxonomy: classify which failures are model issues vs UI issues vs data issues
• Eval writeup: ≥10 evaluation tasks with pass/fail notes and suggested fixes per failure class
• Prompt/config cards: document the settings that work on consumer hardware for common note workflows
• Feedback catalog: own the monthly synthesis of the program’s model-feedback entries

Qualifications

• Currently enrolled student (CS, data science, linguistics, cognitive science, or adjacent — or demonstrated equivalent skill)
• Hands-on comfort with local/offline LLMs or a strong record of learning tooling fast
• Rigorous, honest experimental habits: you catalog failures with the same care as wins
• Clear technical writing — eval writeups are the hard currency of the program
• Comfort measuring on modest consumer hardware and reporting real numbers
• Basic spreadsheet/data-organization skills for catalogs and taxonomies

Preferred

• Built or ran local model setups (Ollama, llama.cpp, MLX, LM Studio)
• Eval or benchmark coursework or personal projects
• Prompt engineering or model-behavior analysis experience
• Familiarity with model formats/quantization (GGUF, 4-bit, etc.)
• Published model-feedback or issue-catalog style writing

COTID AI — Internship Program
Individual, campus-based, self-paced internship running 3–6 months. You claim one element of the product and one role, use COTID free on your real coursework for the full program, and every finding you ship feeds the program — your campus work is the work.
Work

• Core loop: real usage → captured evidence → improvement brief → feedback report on your element
• Deliverables reviewed asynchronously by a mentor against a shared rubric
• Your discipline’s lens makes your findings different from every other intern’s

What you earn

• Free COTID for the full 3–6 month program
• Certificate of completion
• Opportunity to continue into an official cohort program
• Not continuing into the cohort? Option to purchase COTID at a discount at the end of the program
• Portfolio-grade deliverables: an eval writeup (≥10 tasks), a prompt/config card set, and a monthly model-feedback synthesis