Embedded/IoT engineer needed for a short assessment (in-fridge/pantry camera device)

What we're building: We're an Australian consumer app company building a household food management product. The next phase is a small camera device that sits inside a fridge or pantry and periodically captures images so the app can track inventory automatically. No manual entry, no barcode scanning. We are about to commission industrial design for the physical object. Before we do that, we need to know what the physics will and won't allow, so the designer isn't drawing something that can't exist. What we need from you: Two single-page constraints documents. Not a full spec, not a build. Just the box the designer has to stay inside for each environment. One for fridge, one for pantry. This shows us the constraint delta and helps us prioritise which we tackle first. For each, address: - Power envelope. What battery options exist, what physical volume do they consume, and what recharge interval is realistic? Rather than us guessing a capture frequency, work backwards: assume a family will tolerate recharging roughly every three months, and tell us what capture frequency that actually supports and what battery volume it costs. If three months is unrealistic, tell us what is. - Camera module and optics. We tested Pi Camera Module v3 and the image quality wasn't good enough for reading product labels at shelf distance. We need high resolution (8MP+) with a very wide field of view (120°+) to capture useful shelf detail from a typical mounting distance. What camera module actually solves this? USB, CSI, IP-based, whatever works. What are the trade-offs in size, power and cost? Assume a target BOM cost of $30 to $60 USD for the whole device at low volume, and flag anywhere the physics forces you above it. - Lighting. IR versus visible LED, and the real component footprint of whichever you recommend. Environmental. For fridge: operating at ~4°C with high humidity, condensation on the lens, cold-weather battery performance, sealing and IP rating implications. For pantry: room temperature, humidity baseline, any thermal or environmental constraints. - Connectivity. For fridge: a closed fridge is close to a Faraday cage. What actually works for getting data out, and what does that add to the device physically? For pantry: realistic options and trade-offs. Note any Australian regulatory implications (ACMA) that would affect cost or timeline. - Mounting. What attachment methods are viable given the weight, the environment, and the need for a family to remove the device for recharging? Magnetic, adhesive, shelf clip, something else. Note anything that rules an option out. - Bounding box. Given all of the above, a rough minimum volume in millimetres for each environment. Deliverable is two pages, one per environment. Plain language where possible, we'll be handing this to a designer who is not an engineer. Where you're estimating rather than certain, say so, and note what you'd need to test to firm it up. We'd rather have honest ranges than false precision. Scope: We estimate 4 to 6 hours. Please cap your proposal accordingly. If you think it needs more, tell us why in your proposal rather than after you start. What this is NOT Not a build or prototype engagement Not a computer vision role. We handle the recognition models. We need your help choosing the sensor and lens that produce a frame worth running them on Not ongoing work, though we expect further hardware sprints later and would rather keep working with someone good! Context that may help We already have a working Raspberry Pi camera rig that captures images successfully. It's a bench prototype only and is not the product. You don't need to work with it, but we can share details if useful. Ideal background Has shipped or prototyped a battery powered camera or sensor device end to end Embedded Linux / Raspberry Pi or equivalent Experience selecting camera modules and lenses, not just integrating them Comfortable writing clearly for a non-engineer audience To apply Please answer both of these briefly: Without recommending a specific part yet, what are the two or three specs that actually decide a camera module for this application, and what would you test before committing? How would you get reliable connectivity out of a closed fridge, and what are the trade-offs of your approach? Also share one example of a battery powered device you've worked on. Confidentiality: An NDA will be sent prior to work commencing.