Asst Prof of ECE @DukeEngineering and CS @DukeTrinity, Postdoc @YaleSeas, PhD @EE_ColumbiaSEAS, wireless | optics | mobile | edge | IoT
Durham, NC
Joined September 2018
- Tweets161
- Following647
- Followers593
- Likes765
๐ฅ Our latest open-source project: ๐๐ค๐ข๐ฅ๐ฅ๐๐ ๐๐ ๐๐ ๐๐ง๐ญ๐ฌ ๐๐จ๐ซ ๐๐ฆ๐๐๐๐๐๐ ๐๐ง๐ ๐๐จ๐ ๐๐ฒ๐ฌ๐ญ๐๐ฆ๐ฌ ๐๐๐ฏ๐๐ฅ๐จ๐ฉ๐ฆ๐๐ง๐ญ + ๐๐จ๐-๐๐ค๐ข๐ฅ๐ฅ๐ฌ๐๐๐ง๐๐ก
๐ ๐๐๐ฉ๐๐ซ: lnkd.in/efQsHhMn
๐ ๐๐จ๐๐ & ๐๐๐ฆ๐จ๐ฌ: lnkd.in/eaDBAuge
๐คย We are continuously evolving this ecosystem from hardware to real-world applications, and welcoming feedback and community contributions! (6/n, n=6)
#EmbeddedSystems #IoT #AgenticAI #LLMs #EdgeAI #Skills #Arduino #ESP32 #Nordic #Zephyr
๐ ๐๐๐ฒ ๐๐๐ค๐๐๐ฐ๐๐ฒ๐ฌ
- Structured human-expert skills largely improve success rates without retrieval or long context reasoning.
- One-shot system generation: Agents can generate fully embedded applications that run on real hardware, from sensing to actuation. (5/n)
โจ ๐๐ข๐ ๐ก๐ฅ๐ข๐ ๐ก๐ญ๐ฌ
- ๐๐ฌ๐ช๐ญ๐ญ๐ด-๐ฃ๐ข๐ด๐ฆ๐ฅ ๐๐ ๐ข๐จ๐ฆ๐ฏ๐ต๐ด (๐ฃ๐ฆ๐บ๐ฐ๐ฏ๐ฅ ๐ฑ๐ณ๐ฐ๐ฎ๐ฑ๐ต๐ช๐ฏ๐จ)
- ๐๐ฆ๐ข๐ญ ๐๐๐ ๐ฆ๐ท๐ข๐ญ๐ถ๐ข๐ต๐ช๐ฐ๐ฏ
- ๐๐ฐ๐-๐๐ฌ๐ช๐ญ๐ญ๐ด๐๐ฆ๐ฏ๐ค๐ฉ
3 platforms, 23 peripherals, 42 tasks across 3 difficulty levels, and 378 real-world experiments! (4/n)
๐ฌ ๐๐ฎ๐ซ ๐๐ฉ๐ฉ๐ซ๐จ๐๐๐ก
We introduce a skills-based agentic framework + a hardware-in-the-loop (HIL) benchmark that evaluates AI agents where it actually matters -- on real embedded systems. (3/n)
๐ก ๐๐ก๐ ๐๐ก๐๐ฅ๐ฅ๐๐ง๐ ๐
LLMs are great at generating codeโฆ but in embedded/IoT systems, code that compiles and flashes โ code that actually works on real hardware. Timing, peripherals, wiring, and physical interactions can break AI-generated solutions on hardware. (2/n)
๐ AI agents are transforming the software ecosystem and are now rapidly entering the physical world. But one fundamental question remains: Can they reliably build usable, deployable embedded and IoT systems? (1/n)
Just in time for the new year: 8x @nvidia RTX 6000 Ada GPUs have arrived at our lab @DukeEngineering. Excited to receive an NVIDIA Academic Grant to advance the development of high-fidelity wireless digital twins with precise 3D world models!
@NVIDIAAIDev #NVIDIAGrant
Tingjun Chen retweeted
๐ Excited to announce the S3 Workshop 2024, co-located with @ACMMobiCom , happening next week, November 18th, in Washington, DC!
We have an exceptional lineup @S3Acm
โจ Keynote by Prof. Mahadev Satyanarayanan (CMU), the pioneer of edge computing.โ
โจ Panel 1: Maximizing Your PhD Journey
Panelists: Nishant Mehrotra (Rice), Nazish Naeem (MIT), Laura Dodds (MIT), @unpluggedtanmay (Stony Brook)
โจ Panel 2: Navigating Early Academic Careers
Panelists: @Mallesh__ (Northeastern), @soltanaghai (UIUC), @TingjunChen (Duke), @HEchenoz (UC Berkeley)
โจ Technical Sessions highlighting groundbreaking research in mobile computing, embedded AI, networking, and wearable technologies.
Check out the full program here: s32024.smartnets.yale.edu/
A big thank you to all the amazing organizers and steering committee Carlee Joe-Wong @ghasempourEE @marc0_fi0re @PrernaKhanna97 @jianding17
Donโt miss this opportunity to connect, learn, and innovate!
#ACMMobiCom #S3Workshop #MobileComputing #EmbeddedAI
Itโs great to be back on campus to share our work within the @srcorg @DARPA CUbiC center and catching up with friends!
EE Professor Keren Bergman and EE alumni @TingjunChen (MS EE โ18, PhD EE โ20) spoke about connectivity networks and systems at the annual CUbiC conference.
Tingjun Chen retweeted
Carlos Enrique Caicedo Bastidas
@iSchoolSU presented our results on co-existence of #PAWRCOSMOS mmWave arrays and @NOAA CESSRST radiometers at the @NSF NRDZ (pronounced โnerdsโโฆ) Spectrum Week
@SpectrumXCenter #NSFFunded @abhiadhikari3 โ at Arlington, VA
๐ฐ Day 4 of the 2024 @NSF Spectrum Week has started! Today brings NSF NDRZ into the agenda - bringing different focuses, research, and innovators to the conversations had on-site.
๐ Learn more about the agenda for the rest of the day at: spectrumweek.org/
ALT The middle of the image is a hexagon that says Day 4: NSF Spectrum Week. There are photos of speakers, panels, and the career fair, that bridge from Day 3 to the morning of Day 4. At the top of the image are the six organizations - NSF SpectrumX, NRDZCOM4, SWIFT, NSMA, WSRD WG, and IEEE DySPAN.
Check out our latest work on mmWave automotive radar vulnerability published at this yearโs @NDSSSymposium, including the videos on the real-time experiments!
Today at @NDSSSymposium, my student David Hunt will present MadRadar, a physical-layer black-box framework for attacks on mmWave FMCW automotive radars, making valid objects disappear (as in the video below), moving their detections, or even insert new 'spoofed' objects.
We are hiring assistant professors across multiple areas in our ECE department! @DukeEngineering @DukeU
Duke ECE is looking to hire multiple Assistant Professors:
academicjobsonline.org/ajo/jโฆ
Excited to share our latest article describing the design and deployment of programmable 28 GHz radios in the #PAWRCOSMOS testbed in collaboration with @IBMResearch, @EE_ColumbiaSEAS, @RutgersSOE, @nyutandon. Looking forward to outdoor 5G mmWave experiments in the near future!
Excited to part of this amazing transcontinental team to work on sustainable and intelligent AI + 6G networks! @TheAthenaInst #PAWRCOSMOS
We're excited @connect_ie to be coordinating 6G-XCEL: 6G- Trans-Continental Edge Learning, just selected for funding in 6G SNS, which brings together researchers across the EU and US to cooperate on building a secure and sustainable foundation for AI for 6G.
Tingjun Chen retweeted
Thanks to @SpectrumXCenter for the opportunity to present our recently-#NSFFunded NRDZ project in the #NSFSpectrumWeek.
The project leverages both #PAWRCOSMOS @CUSEAS & @NOAA CESSRST @CityCollegeNY.
Full presentation by the entire team ๐
youtu.be/CC4ZM9LdR9Q
Excited about this series of workshops w/ Inspiring Minds, NC, and it's the first time that we bring the #PAWRCOSMOS education toolkit outside the NY area! Thanks Korakis group @nyutandon & @EE_ColumbiaSEAS for the support, and kudos to all the students & @TheAthenaInst members!
This Post is from an account that no longer exists. Learn more
Today @ofcconference, we present a transfer learning-based method for accurate EDFA wavelength-dependent gain spectrum prediction, using a comprehensive dataset collected using ROADMs deployed in the #PAWRCOSMOS testbed. Dataset is available on GitHub at bit.ly/424DlXt !
Today at @ofcconference we will present our work on the field trial of real-time, coexisting fiber sensing and 400G transmissions using the #PAWRCOSMOS testbed with dark fiber provided by @CrownCastle and @ZenFi_Networks. Joint work with @NECLabsAmerica, @DanKilper, @gil_zussman
See the Field Trial of Coexistence and Simultaneous Switching of Real-Time Fiber Sensing and 400GbE Supporting DCI and 5G Mobile Services (W3H.4) (A Top Score Paper) with Yue-Kai Huang of NEC Laboratories America on March 8 from 15:00 - 15:15. #OFC23 buff.ly/2FvoQBx
