I currently work as a security analyst. I hold a master’s degree in information security and have been pursuing research in medical device security.
See the nav bar on the left to look through my work for How To Biomanufacture Almost Anything (Fall 2026)!
HTBAA Homework #1: Hardware-Enforced Zero-Trust Security Gateway for Open-Source Artificial Pancreas Systems
(1) What would you like to biomanufacture, and why? Target Output: A physical, low-cost Hardware Zero-Trust Security Gateway designed for DIY Artificial Pancreas Systems (such as OpenAPS and Nightscout). Motivation: In open-source closed-loop medical tech, control algorithms act as the biomanufacturing “process controller,” dynamically calculating real-time insulin dosing based on continuous glucose monitor (CGM) telemetry. However, DIY systems often communicate over unencrypted webhooks or local APIs with static keys. An attacker or corrupted network packet could trigger an unsafe over-delivery of insulin. Building an inline hardware interceptor ensures every dosing command is physically and cryptographically verified before reaching the insulin pump driver.
(2) What living system would you use?
Chassis: Recombinant Escherichia coli (K-12 strains used commercially to synthesise fast-acting human insulin analogues like lispro).
Inputs: Synthetic expression vectors (plasmid DNA), glucose/media, and IPTG for inducible promoter expression. Biological Process: Intracellular expression of proinsulin in inclusion bodies, followed by cell lysis, solubilisation, purification, and enzymatic cleavage into active insulin analogues.
Outputs: Purified recombinant human insulin delivered via automated closed-loop subcutaneous infusion. System Context: While commercial bioreactors synthesise the raw protein, this project treats the human body + automated insulin pump as the downstream biological execution system.
(3) How would you use automation? Automation Platform: Low-cost, dedicated single-board hardware (e.g., an ESP32 or Raspberry Pi Zero) sitting inline between the APS controller and the pump driver. Workflow Automated: Automating real-time telemetry validation and command verification between the CGM, the closed-loop algorithm, and the insulin pump driver. Security Automation: The gateway hardware physically intercepts incoming BLE/Wi-Fi dosing instructions. It automatically parses JSON payloads, verifies HMAC-SHA256 signatures, checks strict physiological delivery limits (preventing accidental mass-dosing glitches), and logs audit trails without human intervention.
(4) How much would you need to make, and how long would it take?
(a) How much?
Scale Target: Securing 1 individual user’s daily automated insulin dosing pipeline for 1 year. Dosing Telemetry
Scale: Continuous Glucose Monitor (CGM) readings (every 5 minutes) = 288 \text{ reads/day} \approx 105,120 \text{ data points/year}. Automated basal adjustments calculated every 5 minutes = 105,120 \text{ automated dosing commands/year}. Total security payload verifications required = 105,120 API transactions verified per year for 1 person.
(b) How long?
Execution Throughput: Hardware-accelerated cryptographic signature verification on an ESP32 takes < 15 \text{ milliseconds} per request.
Time Estimate: To process a full year’s worth of real-time dosing checks (105,120 loops), the device runs continuously, processing each transaction in less than 0.02 seconds without adding latency to time-sensitive dosing cycles.
(5) What don’t you know yet?
Hardware Power Efficiency: Optimising sleep cycles and power consumption on the ESP32/Raspberry Pi Zero so the security hardware doesn’t drain its battery pack midday.
Fail-Safe Fallback Behaviour: How the hardware should handle edge-case network dropouts or rejected signatures so the pump safely defaults to standard background basal delivery rather than halting insulin delivery entirely.
General Practical & Process Unknowns (“Unknown Unknowns”): As this is my first time embarking on an independent biomanufacturing software & hardware integration project, I recognise that I will inevitably encounter unforeseen technical bottlenecks, protocol edge cases, and hardware bugs during development.
Out-of-Pocket Hardware Prototype Budget Microcontroller (ESP32 / Raspberry Pi Zero W):
~$15 – $25 LiPo Battery + Charging Board: ~$10 Enclosure / Cables: ~$5 Total Prototype Cost: ~$30 – $40