Silicon Valley · Summer 2026 · one radar, 65 conversations, one direction

I brought two things to the Valley:
a working radar, and the business it taught me.

I'm a MIMO radar engineer. I came to the Valley with a radar you can watch run and a business hypothesis to test. Six weeks and 65 conversations later, the hypothesis is dead, and something better is on the table: real demand from the people who build unmanned platforms.
THE TWO THINGS

One. A working X-band phased-array radar prototype I built entirely myself, from the antenna array through RFSoC signal processing to real-time tracking: 7 kg, 2 km detection range, ±30° electronic scanning, no mechanical rotation.

Two. The business it taught me, after the market corrected it once: I arrived pitching per-unit licensed calibration software. Sixty-five conversations killed that idea and surfaced real demand instead: a Taiwanese unmanned-surface-vessel maker heard the spec and said "this is the one we need." The next conversation is about integration on board.

The ambition, one step at a time: from one radar on one vessel to a decentralized network of companion radars along Taiwan's coast.

1 working prototypeX-band MIMO radar: antenna array, RFSoC signal processing, real-time tracking
65 conversationsthey killed the calibration idea, and surfaced real demand from a USV maker
2 asksintros to Taiwanese platform makers; a design partner for a paid PoC
Now · August 2026
  • Valley chapter closed: pitched on Aug 17, program wrapped Aug 21, back in Taiwan.
  • Back in Taiwan next: continuing with the USV maker on integration and PoC terms.
  • By year end: the v2 X-band prototype, ground demo detecting USVs and drones; then weight cut toward a 5 kg tethered-drone payload.
  • Ongoing: the master's thesis behind this radar, and lining up the second and third platform-maker interviews.
Updated 2026-09-05
// THE BUSINESS IDEA, TESTED

I brought two hypotheses. The market kept one.

Six weeks, 65 conversations, three rounds of deep research. The calibration-licensing idea is dead. Selling directly into the US defense market is dead. What survived: a radar module that a Taiwanese unmanned-surface-vessel maker says they need, and a follow-up conversation about integration already on the calendar.
The direction now
  • Sell radar modules to platform makers: USV builders, drone makers, system integrators. They integrate, they face the military or fishery buyer; I never sell defense directly.
  • Fast-updatable algorithms (against jamming and countermeasures) are both a technical requirement and the selling point.
  • Taiwan's chain is complete for this: antenna manufacturing plus system integration, next door to the platform makers.
The business model
  • Staged NRE per platform (US$100K–250K, the structure I priced out in the Valley), then a per-module fee as the platform ships.
  • SBIR Phase 2 (capped around NT$12M over two years) can cover most of a platform maker's NRE: the buyer's risk drops to near zero.
  • Priced for Taiwan: the pitch is "how much each shipped unit saves you," not a vision slide.
Why now
  • Taiwan's navy is about to buy USVs at scale: the founder cited a tender of 1,300 to 1,500 vessels (his words, directional only).
  • The Ukraine lesson hasn't become a US program yet: small companion radars riding on unmanned platforms are still an open slot.
  • Cameras fail at sea in fog, rain and darkness, exactly when detection matters most; radar is the complement platforms are missing.
The honest catch: one conversation is not a market
Three things I hold openly:
  • The traction is a single sample. One founder's "this is the one we need" is a verbal intent, not a signed PoC, and "money is not a problem" is not a negotiated price. The next step is a second and third platform-maker interview, exactly to kill this risk.
  • Bird flocks and buoys are unproven. Small targets at sea are a genuinely hard radar problem. I claim nothing there before the v2 feasibility tests; until then they are scenarios the customer named, not specs I promise.
  • The big numbers are secondhand. The 1,300 to 1,500 vessel tender is the founder's account; I quote it as direction, never as my own market sizing.

Objections I actually got, answered

Sixty-five conversations means sixty-five chances to get attacked. These are the objections that came up most, and where I actually stand.

Three asks now:
  • Intros: technical contacts at Taiwanese drone makers, shipyards and system integrators, for the second and third customer interviews.
  • A design partner: one platform maker willing to run a paid PoC on the v2 prototype.
  • An advisor: someone senior from the defense world who keeps correcting my map.
If that's you, I'd genuinely love to talk.
// MY WORKING MAP OF THE INDUSTRY

Where I think Taiwan sits — tell me where I'm wrong

Two frameworks: pick the battlefield, then decide how to win.
  • Six-stage supply chain: margin is tied to position. The same radar module earns ~20% in contract assembly, far more at the algorithm and IP stages.
  • Where the old idea sat: stage 5. That was my departure map; this summer's research showed the turnkey testers (NOFFZ, Konrad, Averna) already own that ground, so the licensing idea was retired. The full verdict lives on The Business Idea page.
  • Why the map still matters: the surviving direction, radar modules for platform makers, sits at the module and system-integration stages, exactly where Taiwan's chain is complete.

Each stage below is clickable — open one to see its players, its margin logic and where Taiwan stands.

// THE WORKING PROTOTYPE

An X-band MIMO radar, built end-to-end by one person

This is the first of the two things: a working X-band MIMO radar prototype I built entirely myself:
  • the antenna array and RF front-end
  • the RFSoC (ZCU111) signal-processing chain
  • the real-time tracking on the host
It's suited for drone detection and perimeter sensing. I'm demoing it this summer and looking for design partners. The simulation below is a live version of its signal chain: move the controls and everything recomputes.
Lab bench photo: a ZCU111 RFSoC development board wired to the antenna front-end and power boards.
The actual bench: a ZCU111 RFSoC (right) driving the antenna front-end and power boards (left). A photo, not a render.
7 kgtotal weight
2 kmdetection range
±30°2-D electronic scan
X-bandphased array, no mechanical rotation

The first-generation S-band version has already detected drones in live ground tests. Everything below is a browser simulation of this exact signal chain; the photo is the real thing.

New to radar? In plain terms: the glowing wedge is where the antenna is "looking." Steer it across a target and watch the echo pop out of the noise (the range–Doppler map) and get tracked on the scope.

Try it: click an object in the chart below to drop it into the radar above — pick the F-16 (clearly detected), then the F-117 stealth jet: a huge plane that nearly vanishes. The radar shows whether it was detected; raise SNR or add TX/RX elements to dig the faint ones out.

Tracking algorithm
PPI SCOPE · live
System chain
BEAM PATTERN — |AF(θ)|
RANGE–DOPPLER MAP + CFAR
Radar cross-section (RCS) — why a stealth jet reflects less than a bird

RCS is set by shape & materials, not size. The shaded objects are stealth aircraft — huge, yet they return a smaller echo than a bird; an F-16 returns ~1000× more. Click any object to drop it into the radar above and see if it can be detected.

// THE 99-DAY FIELD LOG

Two hypotheses walked in. One walked out.

99 days, 65 conversations. I brought the radar prototype and two business hypotheses to the Valley and let the market correct me, one conversation at a time. Every blip below is one conversation: the angle is time, and the closer to the center, the deeper the talk.

This field log is private.

Enter the passphrase to unlock the 65-conversation log.

THE ENGINEER ASKING

A bit about me

Robert Chang — MIMO radar engineer and the researcher behind this project
I'm Robert Chang, a master's student at NTU's Graduate Institute of Communication Engineering.
  • I design and validate information-fusion algorithms for MIMO radar.
  • I came up through NTHU EECS on terahertz imaging and signal processing — comfortable from the RF bench to the FPGA to the host-side algorithm.
  • My advisor founded a Taiwanese startup doing antenna design, RF front-ends and satellite communications — I build its baseband and back-end, so I'm looking at this supply chain from inside it.
What I enjoy most is being the person who has actually touched every layer of a system, then going to find its real ceiling. Outside the lab:
  • cross-border work bridging technology and business in Epoch Foundation's MIT group;
  • co-founded a non-profit repair community.
Happy to talk radar, dual-use hardware, or anything that connects Taiwan and Silicon Valley.
  • Based inBased in Hsinchu and Taipei, Taiwan; in the Bay Area Jul–Aug 2026
  • EducationM.S. Communication Eng., NTU (2025–27) · B.S. EECS, NTHU (2021–25)
  • Focus nowMIMO radar baseband on a ZCU111 RFSoC — DBF, pulse compression, CFAR, tracking
  • LanguagesMandarin (native) · English (professional)
  • CertificationsAzure AI Fundamentals (Microsoft) · Transformer-Based NLP (NVIDIA DLI) · Altium PCB Design

Built & been part of

Honors & highlights

LET'S GRAB A COFFEE

If you're one of these, let's grab a coffee

I'll be in the Bay Area this summer and I'd love to meet:
  • Jul 6 – Aug 21, coffee's on me — not in the Bay? A video call works just as well.
  • I'm carrying two things: a radar I can demo, and a licensing idea I want you to attack.
  • Design partners get the live demo; skeptics get my honest numbers.
Here's who I'm hoping to meet, and what I'd ask.

Rather reach out now? Email me, connect on LinkedIn, or book a coffee slot — any of the three below works.

Not on this list but think our worlds overlap? Reach out anyway — coffee's still on me.

Robert Chang· NTU Communication Engineering· Global Innovation & Entrepreneurship: Silicon Valley Practicum