<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Carla Prados Bodega on Carla Prados Bodega — Biomedical Engineering student</title><link>https://carlaprados.com/</link><description>Recent content in Carla Prados Bodega on Carla Prados Bodega — Biomedical Engineering student</description><generator>Hugo</generator><language>en-GB</language><lastBuildDate>Fri, 14 Aug 2026 00:00:00 +0100</lastBuildDate><atom:link href="https://carlaprados.com/index.xml" rel="self" type="application/rss+xml"/><item><title>Characterising a steerable fibre under curvature</title><link>https://carlaprados.com/projects/fibre-characterisation-campaign/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/fibre-characterisation-campaign/</guid><description>Six fibres bent past their rated limits, 285 recordings and a 405,000-row dataset. The 200 µm fibre reached the surgical bend and kept 92% of its light there, against 82% for the fibre used today.</description></item><item><title>Sizing a sand battery for 17 Spanish houses</title><link>https://carlaprados.com/lab/sand-battery/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/lab/sand-battery/</guid><description>Chapter 4 of an ENGF0004 coursework, rebuilt as an executable model — every hand calculation becomes a function, and every result is checked against the figure printed in the report.</description></item><item><title>Undergraduate Researcher</title><link>https://carlaprados.com/experience/misi-research-studentship/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/experience/misi-research-studentship/</guid><description>I&amp;rsquo;m running 312 tests on optical fibre under tight bending and building the pipeline that processes the results.</description></item><item><title>An African mask, drawn with a turtle</title><link>https://carlaprados.com/lab/turtle-african-mask/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/lab/turtle-african-mask/</guid><description>The oldest code I still have. I was eleven, and I had worked out how to draw a curve using only straight lines — which turns out to be the whole idea behind how computers draw curves at all.</description></item><item><title>Electromedicine Intern</title><link>https://carlaprados.com/experience/gee-electromedicine/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/experience/gee-electromedicine/</guid><description>I diagnosed and repaired hospital equipment, including ventilators, anaesthesia machines, dialysis units, defibrillators and neonatal incubators. This included soldering and component-level repairs, and I wrote the technical reports.</description></item><item><title>Synthetic bone analogues under load</title><link>https://carlaprados.com/projects/synthetic-bone-analogues/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/synthetic-bone-analogues/</guid><description>Mechanical testing of synthetic bone samples, with more than 50,000 data points per sample analysed in MATLAB. We carried instrument uncertainty through to every result and found three systematic reasons our values differed from the literature.</description></item><item><title>Cybersecurity and Behaviour Analysis Intern</title><link>https://carlaprados.com/experience/amplia-cybersecurity/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/experience/amplia-cybersecurity/</guid><description>I helped the team validate and explain anomaly-detection models (Isolation Forests and autoencoders) on industrial network traffic. There were no labels for normal traffic, so validation relied on consistency and stability checks rather than accuracy.</description></item><item><title>MyLenoma: a skin-cancer triage app in a week</title><link>https://carlaprados.com/projects/mylenoma-triage-app/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/mylenoma-triage-app/</guid><description>We retrained a ResNet50 on five lesion classes using transfer learning, added Grad-CAM to show what the model was looking at, and built the SQL database behind the app, all in one week.</description></item><item><title>Optical ultrasound for lung-cancer diagnosis</title><link>https://carlaprados.com/projects/optical-ultrasound-business-case/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/optical-ultrasound-business-case/</guid><description>A business case for an optical ultrasound venture that doesn&amp;rsquo;t have revenue yet: the size of the European market, competitors and IP, and what CE Class IIa approval would involve.</description></item><item><title>R&amp;D and Quality Control Intern</title><link>https://carlaprados.com/experience/grupo-juste-rd/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/experience/grupo-juste-rd/</guid><description>I worked across R&amp;amp;D and Quality Control at a regulated pharmaceutical manufacturer, running HPLC, gas chromatography and AOX analysis.</description></item><item><title>Safer late-night bus journeys in Outer London</title><link>https://carlaprados.com/projects/tfl-transport-intervention/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/tfl-transport-intervention/</guid><description>Our How to Change the World project: a request-stop scheme letting passengers get off between stops after 10 pm, adapted from Vigo, Spain. It won the 2026 award, and TfL invited us to present it at its National Engineering Day.</description></item><item><title>Where a femur breaks in a fall</title><link>https://carlaprados.com/projects/femoral-fracture-fea/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/femoral-fracture-fea/</guid><description>A finite-element model of a patient-specific femur under impact, tested at different angles and boundary conditions and checked against fracture patterns seen in clinical practice. Graded 100%.</description></item><item><title>An ECG built from discrete components</title><link>https://carlaprados.com/projects/ecg-acquisition-system/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/ecg-acquisition-system/</guid><description>A full biopotential acquisition chain — instrumentation amplifier, multi-stage filtering, 600 Hz sampling — built on breadboard and tested against a clinical database. Graded 92%.</description></item><item><title>An EMG-controlled mouse, built in a week</title><link>https://carlaprados.com/projects/emg-controlled-mouse/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/emg-controlled-mouse/</guid><description>A five-channel surface-EMG interface letting someone with limited mobility drive a cursor. Specified from signal physics, delivered in one week. Graded 91%.</description></item><item><title>A vaccine bioreactor for Uganda</title><link>https://carlaprados.com/projects/vaccine-bioreactor-uganda/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/vaccine-bioreactor-uganda/</guid><description>A small-scale yeast bioreactor to de-risk investment in a full-scale TB vaccine plant, integrating three engineering disciplines&amp;rsquo; workstreams into one design.</description></item><item><title>Histotripsy versus thermal ablation</title><link>https://carlaprados.com/projects/histotripsy-vs-thermal-ablation/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/histotripsy-vs-thermal-ablation/</guid><description>A comparative assessment of five tumour-ablation modalities, reasoning from physical mechanism to clinical consequence — plus a dedicated ethics report.</description></item><item><title>What a 70% price cut did to Eli Lilly</title><link>https://carlaprados.com/projects/eli-lilly-insulin-pricing/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/eli-lilly-insulin-pricing/</guid><description>Analysing a listed pharmaceutical company&amp;rsquo;s voluntary 70% price cut against its own product-level revenue reports and Refinitiv share-price data around the announcement.</description></item><item><title>A peristaltic pump, taken to manufacturable drawings</title><link>https://carlaprados.com/projects/peristaltic-pump-dfm/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/peristaltic-pump-dfm/</guid><description>A full design-for-manufacture package — gear train, casing, connectors — as a complete technical drawing set with tolerances, fits and written manufacturing instructions.</description></item><item><title>An LED keychain in KiCad</title><link>https://carlaprados.com/projects/led-keychain-pcb/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/led-keychain-pcb/</guid><description>Schematic capture and PCB layout for a small LED keychain — a self-directed board design outside any coursework, iterated across five revisions in a day.</description></item><item><title>A wearable infant safety monitor</title><link>https://carlaprados.com/projects/wearable-infant-safety-system/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/wearable-infant-safety-system/</guid><description>Temperature and pressure sensing with embedded decision logic to detect hazardous positions and overheating — designed to keep false alarms low.</description></item><item><title>A living aid, and instructions anyone can follow</title><link>https://carlaprados.com/projects/living-aid-easy-read/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/living-aid-easy-read/</guid><description>Designing the device was half the work. The other half was documentation written to formal easy-read guidance, for users with cognitive or literacy barriers.</description></item><item><title>A menstrual cup as a biomedical device</title><link>https://carlaprados.com/projects/menstrual-cup-device-analysis/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/menstrual-cup-device-analysis/</guid><description>Comparative analysis of cup variants against the underlying physiology, resolved into a decision flowchart a user could actually follow.</description></item><item><title>Plug and socket assemblies in SolidWorks</title><link>https://carlaprados.com/projects/cad-plug-socket-assemblies/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/cad-plug-socket-assemblies/</guid><description>Two size variants — 2 mm and 4 mm — taken from part model to formal technical drawings.</description></item><item><title>Applied tissue engineering</title><link>https://carlaprados.com/projects/tissue-engineering-coursework/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://carlaprados.com/projects/tissue-engineering-coursework/</guid><description>Group coursework on scaffold design for directing cell behaviour, characterisation methods, and decellularisation as a route to whole-organ regeneration.</description></item><item><title>Three sources, one frame rate, and a 35× disagreement</title><link>https://carlaprados.com/blog/three-sources-one-frame-rate/</link><pubDate>Fri, 14 Aug 2026 00:00:00 +0100</pubDate><guid>https://carlaprados.com/blog/three-sources-one-frame-rate/</guid><description>Every video in my dataset reported its frame rate three times, and the three readings did not agree. Resolving that was most of the project.</description></item><item><title>Publishing where your data fails</title><link>https://carlaprados.com/blog/publishing-where-your-data-fails/</link><pubDate>Thu, 02 Jul 2026 00:00:00 +0100</pubDate><guid>https://carlaprados.com/blog/publishing-where-your-data-fails/</guid><description>Tracking failed on some runs. Instead of averaging over them, I mapped where and why it failed.</description></item><item><title>What a 50%-accurate model taught me</title><link>https://carlaprados.com/blog/what-a-fifty-percent-model-taught-me/</link><pubDate>Wed, 20 May 2026 00:00:00 +0100</pubDate><guid>https://carlaprados.com/blog/what-a-fifty-percent-model-taught-me/</guid><description>We built a skin-cancer triage app in a week and won a presentation award for it. The model was about 50% accurate. Both of those facts belong in the write-up.</description></item></channel></rss>