🇺🇸 Samvith Mahadevan

What the device does

The Society for Science describes the project as a machine-learning-trained chemical "nose" that detects allergens such as nuts and eggs in food products, including processed foods. The phrase is a term of art: an electronic nose is an array of sensors whose combined response to a sample forms a pattern, which software then classifies.

The machine-learning element is what makes that workable. No single sensor identifies an allergen on its own; the identification comes from the shape of the whole response across the array, which is exactly the kind of pattern a trained model handles better than a fixed rule.

Training such a system means collecting responses from known samples until the model can separate the patterns reliably. That data-gathering, rather than the electronics, is usually the laborious part of an electronic-nose project, and it is the step that distinguishes a working classifier from a demonstration.

Why processed food is the hard case

Whole ingredients are the easy problem. Processed food is where allergen testing gets difficult, because the allergen may be present in traces, transformed by cooking, or introduced by shared equipment rather than by the recipe. Labelling covers intent; it cannot always cover contamination.

A portable detector attacks the residual risk that a label leaves behind. For families managing a serious food allergy, the practical question at a restaurant table or in front of an unfamiliar packet is not what the ingredient list claims but what is actually in the food.

The competition's own framing helps here. The Junior Innovators Challenge asks middle-school students to identify a problem and build something that addresses it, and judges assess the reasoning and the collaboration as much as the object. A detector aimed at a hazard the entrant can describe from daily life fits that brief precisely.

The award

The Lemelson Foundation Award for Invention carries $10,000 and is one of the named prizes at the Junior Innovators Challenge, the national middle-school science competition run by Society for Science. Mahadevan received it at the ceremony in Washington, D.C. on 29 October 2024.

The top award that year, the $25,000 Thermo Fisher Scientific ASCEND Award, went to Tina Jin, 13, of San Jose, California, for a water filtration system made from animal bone whose output passed potable-water testing by the San Jose Water Company.

The competition

Thirty finalists were selected from nearly 2,000 applicants, drawn from 48 states along with American Samoa, Guam, the Northern Mariana Islands, Puerto Rico and the US Virgin Islands. The programme reaches roughly 65,000 middle-school students a year, and $100,000 in prizes was distributed in 2024.

The other $10,000 awards went to Gary Allen Montelongo of La Joya, Texas, for train suspension modelling, Sophie Tong of Palo Alto for algorithms improving vehicle vision in fog and darkness, and Tyler Malkin of Greenwich, Connecticut, for a saliva test for iodine deficiency.

What is documented — and what is not

Firm: the award, the amount, his age, his home city, the nature of the device and the allergens named, and the date and place of the ceremony.

Open: the sensors used, the accuracy, the detection limits, the cost, and whether the device has been developed since. Samvith Mahadevan is a child; this page carries only what the organisers published, and makes no claim that the detector is available or validated for safety-critical use.