Stalking detection kit
Live on the App Store and revenue-generating.
An iPhone toolkit for finding hidden cameras and trackers. The one that pays for the others.
Swift
London · 51.46°N · Running a family office portfolio · building tools
Dr MikeSemenov
Operator and investor with a constant itch to build.
I run investments for a family office: built its brand and the venture, private equity and private credit deal flow behind it, from sourcing to due diligence, with the back office automated. Angel cheques into deeptech through Material Investments since 2023. Before that, four years across Deutsche Bank's technology, data, derivatives and advisory desks, and a PhD in molecular physics at UCL. I still write the code: nine things built, two live, one earning.
01 / Projects
Built, mostly for the fun of it.
A couple live, the rest tried, shipped and shelved. Some with friends. The constant: small, cheap to run, and fun to make.
Duo
Open source and in use; it produced the PN, SiN and AlH line lists in my papers.
The ExoMol group's Fortran program for computing rovibronic spectra of diatomic molecules. I contributed during my PhD, including the Landé g-factor work. Fortran is still what the physics runs on.
Fortran
VisitAutomatic crypto trading
Running with real money; verdict on the results to come.
Automated trading strategies, backtested and then let loose on a small book. Partly an excuse to build the pipeline.
Family office operating system
Internal; runs the day-to-day for more than one family office.
One system across family offices: CRM, deal flow and document automation, KYC and compliance checks. AI reads and drafts in the background; a human still signs. Built because the off-the-shelf tools assume a single office with a big back office.
trendi
Shelved after eighteen months: Europe's VC market was not ready, and the model does not scale like a venture business.
Access to venture funds at smaller ticket sizes, along the lines of Allocate in the US. Co-founded, built, and shut down on purpose rather than left to drift.
Patchleads
Shelved: no market fit.
B2B lead feeds built from UK public registers, sold as subscriptions. The data pipeline worked; the buyers did not turn up.
02 / Experience
Bank, lab, start-ups, cheque book.
- Nov 2023 — now
Angel Investor · Material Investments INVEST
Early-stage deeptech, FFF to pre-seed, occasional later rounds.
- Jun 2023 — now
Investment Director (VC, PE & RE) · Family Office INVEST
Built the office's investment brand from scratch and the venture, private equity, private credit and real estate deal flow behind it: sourcing, due diligence and portfolio management across the US and Europe, with the back office automated.
- Sep 2021 — now
Geek in Residence · Decacorn Capital INVEST
Technical due diligence on deeptech and AI investments.
- Jan 2020 — Jul 2024
PhD, Theoretical Molecular Physics (with applications of machine learning) · University College London EDU
ExoMol group (Tennyson, Yurchenko). Variational line lists for PN, SiN and AlH used to read exoplanet atmospheres. Nine papers, four as first author.
- Jul 2022 — Jan 2024
Co-Founder · trendi
Co-founded a platform giving individuals access to VC funds at smaller tickets, like Allocate in the US. Shut it down after eighteen months: the model does not scale like a venture business, and Europe's VC ecosystem was not ready.
- Jul 2022 — Jul 2023
Investment Analyst · Collective Equity INVEST
Due diligence on start-ups; advised partners on Fund 1 selection.
- Sep 2017 — Jan 2020
Lead Business Analyst, Origination & Advisory · Deutsche Bank
Led BA effort across teams on a client-outreach platform; MiFID II, research unbundling and Brexit client documentation; CRM data model.
- Mar 2017 — Sep 2017
Java Developer & Business Analyst, Listed Derivatives · Deutsche Bank
MiFID II features in the order management system.
earlier → 6 more
- Aug 2016 — Mar 2017
Business Analyst & Salesforce Developer, CDO · Deutsche Bank
Data models, BRD/FRD, stakeholder workshops, Salesforce modules.
- 2012 — 2016
MSci Physics · University College London EDU
- Jun 2015 — Sep 2015
Summer Intern Analyst, GTO · Deutsche Bank
FX and credit clearing applications; restructuring projects.
- Jun 2014 — Sep 2014
Communications Analyst & Web Developer · Capital List (Capital Enterprise)
Built capitallist.co; start-up and VC outreach.
- Jul 2011 — Jul 2011
Research Assistant · Max Planck Institute for Extraterrestrial Physics
Plasma physics experiments; Fortran code analysing the Collatz conjecture.
- 2010 — 2012
A-Levels: Maths, Further Maths, Computing, Physics · City of London Freemen's School EDU
Silver, British Physics Olympiad.
03 / Research
Line lists for exoplanet atmospheres.
A line list is the catalogue of light frequencies a molecule absorbs. Telescopes match them against starlight filtered through a planet's atmosphere to say what it is made of. In UCL's ExoMol group I computed them for PN, SiN and AlH from first principles and refined them against laboratory spectra.
First author (04)
- 2024
ExoMol line lists – LXIV. Empirical rovibronic spectra of phosphorous mononitride (PN) covering the IR and UV regions
Monthly Notices of the Royal Astronomical Society 536(1), 714–727
- 2022
ExoMol line lists – XLVI. Empirical rovibronic spectra of silicon mononitrate (SiN) covering the six lowest electronic states and four isotopologues
Monthly Notices of the Royal Astronomical Society 516(1), 1158–1169
- 2021
Rovibronic spectroscopy of PN from first principles
Physical Chemistry Chemical Physics 23(38), 22057–22066
- 2016
Predicted Landé g-factors for open shell diatomic molecules
Journal of Molecular Spectroscopy 330, 57–62
Co-author (05)
- 2024
The 2024 release of the ExoMol database: Molecular line lists for exoplanet and other hot atmospheres
Journal of Quantitative Spectroscopy and Radiative Transfer 326, 109083
- 2023
Improved potential-energy and dipole-moment functions of the ground electronic state of phosphorus nitride
Journal of Molecular Spectroscopy 395, 111804
- 2023
ExoMol line lists – LIV. Empirical line lists for AlH and AlD and experimental emission spectroscopy of AlD in A¹Π (v = 0, 1, 2)
Monthly Notices of the Royal Astronomical Society 527(4), 9736–9756
- 2021
ExoMol line lists – XLIV. Infrared and ultraviolet line list for silicon monoxide (28Si16O)
Monthly Notices of the Royal Astronomical Society 510(1), 903–919
- 2020
The 2020 release of the ExoMol database: Molecular line lists for exoplanet and other hot atmospheres
Journal of Quantitative Spectroscopy and Radiative Transfer 255, 107228
Thesis (01)
- 2025
Theoretical studies of diatomic molecules: line lists for Phosphorus Mononitride, Silicon Mononitride and Aluminium Monohydride
Doctoral thesis (PhD), University College London
04 / Theses
Where I think this is going.
Five things I would bet on. Opinions, with reasons.
-
01
Agents will be entities, the way companies are.
A company is a fiction we agreed to treat as a person: it can own, contract, sue and be sued, and someone is liable when it misbehaves. Software agents that transact, hire and hold budgets will need the same wrapper, because the user will not be in the loop and the model provider will not accept the liability. Expect registries, capital requirements and something like directors' duties. The legal form will matter more than the model.
-
02
Space is now an industry, and it is staying.
Launch cost has fallen more than tenfold since the Shuttle and the cadence is weekly, not yearly. That turns orbit from a government programme into a supply chain: manufacturing in microgravity, Earth observation sold by the pixel, communications that bypass terrestrial incumbents. The interesting companies are not building rockets. They are building the boring things that assume the rocket exists: insurance, logistics, data, tooling. That is where the margin was on every previous frontier.
-
03
Millennials will not die of old age.
Not because of one breakthrough but because of compounding ones: ageing treated as a disease with biomarkers rather than a fact of life, gene editing dosed per patient, organs grown or printed, the first drugs that slow biological clocks already in trials. Someone born in 1990 has sixty years for the science to arrive, and it only needs to arrive in stages. Dying of old age will look like dying of smallpox: a thing that used to happen.
-
04
Personalised medicine is already here.
In May 2025 an infant in Philadelphia received a base-editing therapy designed for his single mutation, from diagnosis to dose in about six months. That is not a demo; it is a template. Sequencing is cheap, models such as AlphaGenome predict what each variant does, and manufacturing is shrinking to the scale of one. The bottleneck is regulation and reimbursement built for blockbusters. Whoever makes N-of-1 routine rather than heroic owns the next decade of medicine.
-
05
Quantum communication will be standard infrastructure.
Quantum key distribution already runs on metropolitan fibre in several countries and has been demonstrated by satellite. Once a cryptographically relevant quantum computer is plausible, every bank, government and cloud needs channels that do not depend on hard maths staying hard. That demand does not wait for the computer to exist; it arrives with the insurance premium. Expect it to be bought the way TLS certificates are: invisible, mandatory and boring within ten years.
05 / Signal
Things worth a second look.
Pieces I keep coming back to, with a line on why.
- 8 Sep 2026
Google DeepMind AlphaGenome Atlas: molecular predictions for 9 billion human DNA variants
A precomputed map of what every possible single-letter change in the genome is predicted to do. Biology is becoming a lookup table.
- 21 Jan 2026
Nature Probing quantum mechanics with nanoparticle matter-wave interferometry
Matter-wave interference pushed to nanoparticles: the quantum-to-classical boundary tested with objects you can almost see.
- 15 May 2025
Children's Hospital of Philadelphia World's first patient treated with personalised CRISPR gene-editing therapy
A bespoke base-editing therapy designed, tested and dosed for one infant in about six months. N-of-1 medicine is now real.
- 24 Dec 2012
Wired John McAfee fled to Belize, but he couldn't escape himself
Joshua Davis's long read on the antivirus founder's unravelling in Belize. Still the best piece on what money without structure does to a person.
Standing feeds
- Hacker NewsFor all the good stuff.
- Mr Family OfficeFor tracking what is happening in private wealth.
- BloombergFor what is happening live.
- ReutersFor what is happening live.
- Quanta MagazineFor science and tech news.
- TechCrunchFor start-ups.
06 / Library
What I would recommend.
A curated selection of what I would hand to a friend to read or watch. I read far more than I watch.
Books
Red Mars
Kim Stanley Robinson
First of the Mars trilogy: the first hundred colonists argue about terraforming, politics and who owns a planet. Hard science fiction with real economics.
Tools of Titans
Tim Ferriss
Distilled routines and tactics from a few hundred podcast guests. Best dipped into than read through.
East of Eden
John Steinbeck
Two families in the Salinas Valley and the Cain and Abel story retold. Timshel: thou mayest.
Films & series
The Man from Earth
Richard Schenkman
A professor tells his colleagues he is 14,000 years old, in one room, in real time. Written by Jerome Bixby on his deathbed; no effects, all argument.
Amazon links are affiliate links: as an Amazon Associate I earn from qualifying purchases, at no cost to you.
07 / Off-screen
When the laptop is shut.
Make
- Prototyping and tinkering
- Theoretical research Collatz conjecture and friends.
- Reading and writing
Move
- Running, gym, Spartan races
- Sailing
- Hiking and long-distance walking
Taste
- Cooking
- Coffee tasting
- Wine tasting
08 / Contact
Say hello.
Founders, collaborators, journalists, recruiters: the form takes a minute and I read everything.
Clapham, London, United Kingdom. Built with Astro, hosted on Cloudflare, no cookies, no tracking.