Quote from my notes below:

Scary that my 5th year starts in the Fall. I feel old, bad at my job, and very behind. Shoutout to the me who was pretty sure I could graduate in 4 years max.

Between family emergencies, a series of rejections, and stepping down as CEO of AiroSolve, I’m proud of actively working on my physical and mental health.

Things I accomplished:

1. Published my first first-author paper in my actual field of research (IEEE BIBM short paper)
2. Incorporated AiroSolve as a Delaware C-corp before stepping down as CEO
3. Presented at the AMIA Fall Symposium in Washington DC
4. Attended and contributed to the UCLA digital sensing workshop, meeting leaders in my field
5. Went to Nashville for the MCPsych bootcamp and earned a gold star for my contributions
6. Finally found a therapist through UCLA SHIP
7. Co-authored manuscripts on hyper-oxygenation in ICUs and equity in Biodesign
8. Confirmed August 2024 as my expected graduation date
9. Participated in the UC academic workers strike
10. Confirmed wedding venues for both India and US weddings

Things I thought I’d have done by the end of September:

1. Published more in my field (so many rejections: AIME, AMIA student paper, PSB deadline missed)
2. Gotten my main anomaly detection research to work (instead: lots of null results)
3. Secured an Apple internship or some industry experience
4. Felt less behind on my graduation timeline

Things I want to do next year:

1. Graduate (end of August 2024)
2. Get married twice (India in February, US in Spring)
3. Publish meaningfully in digital mental health
4. Figure out my post-graduation career (targeting a Series B digital health startup)
5. Go on a honeymoon

Not the best year for my self-esteem. My AIME paper was rejected. My AMIA student paper was rejected. I missed the PSB deadline. The anomaly detection methods I’d been working on for years now produced all null results. It got to the point where presenting my work felt like an exercise in explaining why I wasn’t dumb for going down paths that didn’t pan out. It’s hard not to take criticism of research you associate with your identity personally, especially when you’re showing slides of all the things that didn’t work.

Stepping down from AiroSolve was one of the hardest decisions I made. I was overwhelmed and unable to keep track of my PhD topic. I had forgotten papers I’d read and taken detailed notes on just a month prior. By stepping back, I could finally focus, and I started to feel like a scientist again. December brought a family emergency that took me to India for over a month - that was its own saga, but I’m glad I went.

Through all of this, I’m proud I maintained my personal life. I found a therapist. I kept surfing. I asked all 7 of my groomsmen to be groomsmen with custom 3D-printed boxes I sanded and stained myself. I went to Coachella. And somewhere in between rejections and pivots, I started to see a clearer path forward in my research and in what I want to do after I graduate.

Notes

Below are real time, minimally altered notes taken during the past year. They capture my thoughts and feelings as they happened, organized by season and date range.

Fall 2022

Sept. 25th to Dec. 1st, 2022

Life Context

This time period had a whole lot of chaos, both good and bad. I had some really productive weeks on my projects, went to a conference, and travelled a whole lot. We had guests stay over pretty frequently, did so many fun events. I spent a couple days in the bay area, flew to Michigan for a weekend, flew to Washington DC for a week for the AMIA Fall symposium, went to Irvine for 2 days for the BornInCalifornia startup pitch event, and spent almost a week in Hawaii surfing daily. In October, Noodle (my cat) developed feline idiopathic cystitis and had to undergo surgery after an emergency room hospitalization. I’ve been surfing pretty regularly, my Fiancée got me an amazing new surfboard after she proposed to me reigniting my desire to be in water. I also took a hip hop beginners dance class, and finished a 3 month workout program (MAPS Performance). Overall feeling physically very fit and functional. Starting Nov 14th all academic workers at the University of California (system) started striking. This was prefaced with rallies, and other actions I took part in but did not assume a leadership role in. By the time the strike started I felt broadly unaware of all that had gone in to it happening, and was very overwhelmed but decided to dive in and picketted as much as I could in solidarity. The strike is still going on (18th day) as of today. I feel very behind on research and honestly miss doing it.

AiroSolve

AiroSolve is formally a company now (Delaware C-corp), and it feels like we are making progress, although slowly. Applying for SBIR Phase I grant, pitched at different events, and are steadily filling in our gaps. Everything has gone very slow however our team is figuring out how to best work together and its been really nice seeing all the people including those at Anderson who believe in and support us. Getting more and more excited about pursuing this hard. Little frustrated at pace of events with our patents, a lot of delays both on our and the tech dev group side.

Research

Recently I’ve been focusing more on simulation based result generation as a recenter my work on interpretation of mHealth data more than detecting any specific thing. I primarily am looking at how to use anomaly detection methods to find useful changes in mobile health datasets across different timescales. For the projects I’m working on I have been putting a lot of my time in to project management more than actual research, a project manager was brought on at 20% time and was quickly overextended as I offloaded my responsibilities to them. That process of telling the new project manager what I was doing definitely made me appreciate how much I had passively taken on and that I really was doing project management, something I never explicitly thought about. I feel more certain about my time and where my priorities should lie, glad my advisor was able to get a new PM on board to allow me to focus more on research. The strike has made me so antsy to want to get back to research though, I really do like what I do. Starting to feel like I’m falling behind and won’t hit my desire for a spring 2024 graduation. I also wasn’t able to get an internship at Apple this summer and can’t legally work on mobile health stuff anywhere else. Probably is good for me to stay heads down and focus on what I’m doing at UCLA though. Going to a conference made me realize I should have been going way earlier, and could have submitted even my rotation projects as posters to AMIA. Going as a fourth year I felt old, late, and a little blindsided. Great to have the opportunity to connect with what does seem like my professional home.

Winter 2023

Dec 2nd to Feb 4th 2023

Life Context

Chaos continued, I had a family emergency and flew to India Dec 6th until Jan 15th. That was a whole saga, glad I went but it was pretty tough. Got to really connect with a lot of my family in India which was really nice. The strikes over though so worked on my PhD, stepped down at AiroSolve, and am feeling a bit better about my academic status/progress now.

Also planning a wedding in India Feb 24th 2024, and America in March/April so that’s been a lot of work that I’ve procrastinated. Realizing that I have no idea what constitutes an Indian OR American wedding, but my fiancée does and I’m struggling to figure out where to even start. But trying my hardest now. Setting goals to do some planning per week and start the ball rolling.

AiroSolve

One of the hardest decisions I made was to step down. I was overwhelmed and unable to keep track of my PhD topic. It got to the point where I forgot I read a paper that I took detailed notes on just a month or two ago. Could not keep track of all of the business and my PhD topic enough to be an expert. By taking a step back I was hoping we could transition in to a team more able to bring the company forward. We’ve found a new co-founder, the team seems good, and I feel so much better being able to focus recently on my doctoral work. While both are medical, having to know so much about two distinct fields taxed my brain cells too hard. There was also just too much administrative work for me to do, meetings to attend, etc., that were required but took a lot of time.

Research

I wrote a short paper submitted to AI in Medicine 2023! Much of the work is simulation based and I did it on planes and cars in India. Feels great to have a result. This is my first time really doing something publishable in my field of research. I submitted the manuscript Jan 30th, presented at lab meeting Feb 1st, and put together a 3-minute talk for Grad Slam on it Feb 2nd. I’m now going to extend that work for a student paper submission to AMIA due March 8th. Really excited, but these are tight deadlines. Hoping with my renewed time available to focus I’ll be okay.

Also in the manuscript revision stage for work related to equity in the biodesign process and another paper about hyper-oxygenation in UCLA ICUs. Glad those two projects are wrapping up and look good so far.

We’ve been waiting on a high performance compute environment to be ready for use for a while. Instead we have a compilation of stop-gap solutions for incoming analysis needs that has been a struggle to really work with. I’ve personally just avoided using it in favor of simulation work on my macbook air. My computer I SSH in to at lab also has stopped working so getting that fixed by UCLA Health IT now as well. Macbook air has been a computational powerhouse recently.

Wellcome Leap MCPsych Consortium

I’ve gotten back in to some administrative tasks regarding the consortium here. I went to Nashville Jan 24th for a gathering of 1-2 data-folk from each institution to help harmonize our processing pipelines, data types, and learn the system we are to use for uploading everything. It was actually really useful to physically meet the people involved and feel part of a consortium. I also had no idea what my responsibilities in the consortium were, and now feel like I get that better. We have neither data nor models really available but have to upload something by Feb 15th so working with 2 others at UCLA to make that happen. We have a plan for a simple predictive model and data types to get ready, and while a tight deadline, seems doable. A key issue to overcome is being able to upload data from the compute environment given the array of firewalls put up to secure the data.

Feb 5th to March 17th

Life Context

Been in a bit of a mood slump these last couple weeks. Part of it is finally having time and space to really think about all that happened, but also part of it I don’t understand. I feel like life and work are both on the surface going well, but also I’ve had a lot of feelings of general incompetence and overwhelmed-ness. I think this will pass but also know a little too well which questions on a PHQ-9 would raise a yellow flag for me.

Asked all 7 of my groomsmen to be groomsmen, that was a surprising amount of work I inflicted on myself. I think 3D printing fun boxes was a good idea but the hand sanding and staining of the wooden filament took way too long. Oh well, I think they look okay, and glad the groomsmen seemed to like them.

Went on a professional social bender with a workshop-conference at UCLA from Feb 28th to March 2nd, then flew March 2nd night to another retreat/conference until March 5th evening, went to a concert on March 5th night (Polo & Pan – great time). Drove back to the Bay Area to vibe, work, and do little for a week from March 13th to 18th.

Professional Development

Now that I’m not CEO of a medical device company, I need a plan for what I want to do post graduation. I’ve made it a priority to at least monthly talk to someone in a field of interest to me about their career and as a consequence my career. At the Out-of-packet retreat there was a great workshop on career transitions that helped me figure out that working at a digital health startup in roughly Series B stage (~50-100 people) might be ideal. Of course this is all assuming I’ll get to pick my next job vs try get whatever is available.

There was a workshop at UCLA on the future of digital sensing for mental health, bringing together leaders from academia, industry, government, and funders. I had the opportunity to be a note taker for one of the 4 working groups centered on the problem how how data flows in digital sensing studies. Was able to meet so many of the people in my field who’s papers I’ve admired and who’s work I’ve tried to build on, really an amazing networking experience. Made me feel a little more desire to stay in academia after graduating.

Research

My AIME paper got rejected, definitely sucks when something you’ve worked really hard on is just not good enough, I didn’t disagree with the reviewers either. In my mind the bar for a conference short paper was pretty low, and to not have met that with work I’ve kinda centered my identity around was a blow. But I was able to incorporate a lot of that feedback in to the student paper I’m writing for AMIA fall symposium.

I’ve gotten a much more powerful Macbook pro from lab since I was routinely crashing my personal Macbook air and that has been a game changer for running simulations and analyses locally.

Going to hold off on resubmitting my F31 until the August grant deadline as it will be stronger if I use some data from the study funded by Wellcome Leap to show proof of concept.

Wellcome Leap MCPsych Consortium

Still working on analyzing the differences of GeneActiv watches to Apple watch accelerometer data. Surprisingly difficult to query and use data stored in a Cassandra database which is where our Apple watch study data is. Working to improve the model I helped build for the last February deliverable in prep for the end-of-March deliverable.

Spring 2023

March 18th to May 11th

Life Context

I’ve been pretty good about taking care of myself for these last two months. Doing all the strength training and yoga workouts, roughly something 5/7 days a week. Also been trying to up my baseline activity with walks and consistently but slowly losing weight which feels right. I also found a new therapist finally, didn’t know UCLA SHIP covers unlimited sessions with providers outside of UCLA (if they take insurance), gave me a list of therapists and I found one pretty quick.

We’ve had a lot of birthdays in between with 1 or 2 celebrations every week, watched the Mario movie, went to Coachella and car camped (definitely going again so worth it). In between I went home to Santa Clara for a couple days to see my mom who is back from India. Went to Super Nintendo World in Universal studios, and tried to surf when possible, but only went like 3 times in the last 1.5 months.

Professional Development

I had a conversation with a colleague from my program who had just defended their thesis, and they looked over my resume. Gave me some really good insights and I fully revamped it – feeling pretty good about that. I’ve been going to some biodesign events and it has been nice remembering I do kind of have a solid network, was able to send job postings with referrals to friends and feel pretty confident with my place in LA med tech – however, I need to get deeper in to my own fields networks. The digital sensing workshop definitely helped me meet the right people to start that process and I definitely feel very grateful to have been involved there.

Research

After the AIME rejection I put together an updated manuscript on the same topic but with analysis done on real mHealth data from public datasets. This was then submitted as a full student paper to AMIA, hoping to hear back from them soon but pretty happy with the manuscript overall. Also finalized the hyperoxia educational study manuscript which was both submitted and rejected → but we will resubmit to a more specific journal. We are putting the final touches on a manuscript related to equity in Biodesign as well.

Right now I’ve been spending a lot of time looking at my own HealthKit and Whoop datasets in order to build an anomaly detection visualization system that I personally find useful to then serve as inspiration for my next research step. I’ve shown that AD algorithms appear to pick up on something but don’t know exactly what and it seems different per individual. So now I need to see if what they pick up on is related to what is happening in a persons life → going to try relate anomaly detector top features to PHQ/self-report/EMA domain scores using the Wellcome Leap/OPTIMA data.

WL/OPTIMA

Lots of work done to relate digital phenotyping features to self-reports for OPTIMA as part of deliverables. Also other things like dockerizing pipelines and such that the consortium needs, taking a lot of my time but they are funding me.

Found an error in the processing of SensorKit data so for the last month Ethica has been working towards re-processing that data to fill in the almost 60% of data points that appear to be missing.

May 12th to June 4th

Life Context

Starting trying to slowly gain weight now and workouts have felt great as a consequence. However realizing I have been operating on a too little sleep and enabling that with coffee so spending 2 weeks only drinking coffee if it is a social activity (getting coffee to network, or doing a coffee tasting) rather than keeping it as part of my routine/habits. So far I’ve napped a lot more and starting to feel a bit more energetic without the coffee – definitely the most dependent on coffee I’ve been.

Went to Seattle, took the 35 hour coastal train, was an amazing experience, and spent some time with friends and my partner’s brother. Gotten a lot more serious about wedding planning - we are very behind and overwhelmed trying to find a US venue. Things are expensive…

Research

There have been cool opportunities to put together quick abstracts recently which are somewhat related to my research in anomaly detection and mobile health data analysis. Excited to be working on them but also have to now re-center on some of my core research and start planning for the next set of results.

Recently been working with a student who was rotating in our lab and its been really cool seeing how they approach the data in ways different than my intuition would lead me to. Its also been cool seeing them take some of the code on anomaly detection algorithms I wrote, implement it, and extend it. I think it would be really fun to have undergraduate or other students to work with on subprojects of my research.

Professional Development

Went to see Massimo (a medical device company) and it was super cool. Made some good connections with their wearable device team and was impressed with their technology. A lot of cool companies in Irvine.

Summer 2023

June 5th to July 25th

Life Context

I have been getting deeper in to surfing this last month, the addiction has hit. I’m finally almost kind of an intermediate surfer and can very consistently catch waves on my mid-length board. I’ve also been trying to gain weight but because my baseline activity has actually increased with surfing more regularly (2x weekly) on top of 3x weekly lifting, I have increased my calories by 300-500 daily and maintained my weight at around 160. My body feels stronger which is satisfying and lifts have been getting better.

Wedding planning has been a big part of the last couple months. We finally confirmed our American wedding date and venue and signed the contract. Feeling pretty grateful for the large group of close friends I have and the vibrant personal life I get to experience. Been seeing some of my favorite artists live, celebrating life events, tasting so many cool things (food, wine, coffee), and being so occupied that I have to actively make time to be bored and sit with myself outside of my morning meditation.

Research

Its been a lot of rejection, a lot of null results. My AMIA submission was rejected, my AIME submission was rejected, my ACNP submission as part of a panel was rejected. Anomaly detection on multiple datasets have failed and looking back there are such clear signs that it theoretically wouldn’t work. However, I’m someone who tends to not worry about that until a full pipeline is jankily assembled and debugs/refocuses after. I grinded really hard June and July to get results using anomaly detection methods but I pretty thoroughly showed my method does not work. I can find really interesting time points on my own data but they are not strongly related to longer term mental health when applied to other datasets.

Been working to disseminate and get feedback on the outcome of the digital sensing workshop at UCLA. Its been really good to feel connected to the larger research community handling these issues. I’ve now been the only student on calls with some really great researchers in the field and feel grateful for the opportunity. I think the resulting paper will be a big part of my actual dissertation.

Thinking about my dissertation has been grounding. What I’ve done has largely been research tangential. I’ve recently met or presented my research to all my committee members and my results are lackluster. I do feel like I have a path to doing something cool and some insights are starting to crystallize. Its just really not how I thought it would go. Its weird when talking about my results as being null, it often feels like questions imply that I am dumb for going down this path and there was a clear logical reason this didn’t work when in reality we really just didn’t know. I think a lot of that is in my head, but intonation and wording generally kind of feel that way. Really hard to not take criticism of research that I kind of associate my identity and purpose with, especially when I’m presenting null results as null results. I guess its not really common to have so many slides of the stuff that didn’t work…

Pivoted research direction in the short term to see how sensor data may be used to predict item level responses to self-reports related to different aspects of depression combined with model interpretability (SHAP value analysis). During this PhD, I have never worked as hard as I did this and last week. Had several 11+ hour days, worked a weekend, and stayed up til 2:30am last night to get the manuscript to my advisor for feedback. The results I’m using were started on Wednesday evening and I had a lab meeting presentation on Thursday. Manuscript drafted by Monday night. I really do hope this submission to PSB is accepted, but after being rejected at basically every thing I have applied for, I’m not as confident as before.

Scary that my 5th year starts in the Fall. I feel old, bad at my job, and very behind. Shoutout to the me who was pretty sure I could graduate in 4 years max.

July 26th to September 4th

Life Context

This last month-ish has also been pretty jam packed with work, travel, and fun events. I’ve been very actively now reserving my Sunday nights to prepare and reset for the work week and its helped me a lot in being more productive throughout my week. I think I put my body through a little too much strain, surfing multiple days a week, yoga or going out the remaining days. I’ve been strength training much less frequently but need to resume. (average 1.5x per week for the last month). I’m attempting to review how I approach strength training, I find it kind of boring compared to surfing and other forms of activity but do want to still gain muscle mass so will continue. After a 5+ hour surf session on a Saturday followed by a strength yoga session on Sunday, I pulled my neck muscle just making a pour over coffee in the morning on Monday. We have since bought a Theragun while it was on sale and its been really nice so far, I think it will improve my recovery + flexibility in tandem with my other activities.

Wedding planning continues to be stressful, Shar and I dedicate Thursday nights to it, but as a consequence its been hard to find chill time for us to regularly hang out outside of a larger group setting. Starting to more intentionally schedule nights for us to go on dates and just hang out without a stressful task present. In the midst of all this I’ve been proud that at least once a week I’ve set aside time for me to just do nothing and relax.

Research

After grinding more than I normally do, I got my manuscript ready in time to be reviewed quickly by peers. Unfortunately there were some delays in getting necessary revisions and it was not ready in time for the PSB submission, however, I did further revise it and submit to IEEE BIBM 2023 for the August 15th deadline. Proud I got that in, and I think the work is an interesting first foray in to publishing in this space. I’m pretty confident that this one will be accepted, however, can’t get my hopes up. The work was well received by the Wellcome Leap MCPsych consortium and was the basis for much of what we used in our quarterly milestone report. It was nice to hear from others that my code appears well written and documented enough for them to reproduce. I went to Nashville recently for the second MCPsych Bootcamp bringing together data folk that work in the consortium and earned a gold start in recognition of my work. While a little silly, I did actually feel really good about getting some recognition and appreciated that they took the time to do that.

I’ve been working on revisiting anomaly detection. Hindsight is 20/20 and I see that a lot of what didn’t work for anomaly detection is related to how much I tried to do at once. I need to pair down, specify, and establish some baselines to see where anomaly detection fits in to the picture of mental health related monitoring with wearable devices. I’m focusing to begin with on a set of sleep related features and how anomalies relate to changes in sleep quality from questionnaires but also how those features on their own relate to the questionnaires. I’m starting to think more like a scientist and feel like I have a clearer idea of how to move this research direction forward. Its very early stage and experimental which is pretty fun in contrast to the work I did for the IEEE manuscript that was more a direct addition to what has already been done but on new data with slightly different methodologies. I also got preliminary confirmation that the end of August 2024 I’ll likely graduate.

Career

I’m continuing to try and find early stage companies to potentially work at doing digital mental health. I got really lucky and connected to the CTO of one such company and am trying to discover a couple more. Hoping it works but I think my post-graduate career is so surprisingly uncertain at this stage. I always feel like I’ll know what I’m doing in X years before I graduate but now its less than 1 year and I still don’t really know. To be fair though I’m getting kind of closer.