Ankur Kalra

89 posts

Ankur Kalra

Ankur Kalra

@AnkurHop

Katılım Haziran 2018
106 Takip Edilen53 Takipçiler
Ankur Kalra
Ankur Kalra@AnkurHop·
Really appreciating this #DTx Insiders webinar hosted by @decimal_health. "It's not an app, it's a behavioral intervention. Yes it's delivered by software, but that's not the point." (paraphrased version of a key insight from Ameya Phadke of @ChiesiGroup). Love it.
English
0
0
0
110
Ankur Kalra
Ankur Kalra@AnkurHop·
@bnjmnslvr Excited for your Neurohacking class! Do you have a curriculum or agenda or something you can share? Trying to figure out how deep it gets and where it ends up.
English
1
0
0
25
Ankur Kalra
Ankur Kalra@AnkurHop·
Much to my team's detriment, I ultimately hand wrote custom holiday messages myself. It's unfortunate because they clearly deserve the best -- and as we all know, our new robotic AI overlords would be so much better than me. Nonetheless, human hand written cards for now!
English
0
0
1
0
Ankur Kalra
Ankur Kalra@AnkurHop·
4. Buy an @EMSL AxiDraw to neatly yet realistically hand inscribe cards with calligraphy skills that I don't have and can't learn (h/t @aarondfrancis) twitter.com/aarondfrancis/… Okay this might have slightly been because I've always wanted an AxiDraw and needed an excuse.
Aaron Francis@aarondfrancis

It's Friday, so I want to share one of the most fun projects I've ever worked on. A few years ago, I built four eight-foot-long handwriting robots. We wrote many thousands and thousands of addresses on our direct mail, and wrote personal messages for high value prospects.

English
1
0
1
0
Ankur Kalra
Ankur Kalra@AnkurHop·
You've no idea how hard it is for a keyboard-first tech guy to hand write holiday messages. An itemized list of things I semi-seriously considered doing to procrastinate hand inscribing holiday cards for the team:
English
1
0
0
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@nikillinit On it. Haven't read this paper, but familiar with a lot of work in the space. At first glance, the result seems surprising, but will read the paper more closely.
English
1
0
1
0
Ankur Kalra
Ankur Kalra@AnkurHop·
Some great sessions yesterday, but the main program starts this morning! Looking forward to the award paper orals — got an extra large coffee to make sure I can keep up. #CVPR2022
English
0
0
2
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@ihodes some random internet rabbit hole led to your 2014 blog post on NGS -- thank you so much! that was wonderful and illuminating on different levels. as an outsider, reconstructing that from reviewing literature wouldn't have gotten very far. definitely owe you a beer!
English
1
0
0
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@lisa_van_gelder I think that’s the exact challenge — successful exploratory projects don’t obviously look different externally than unsuccessful ones (pre critical breakthrough). We’re currently exploring a crit based approach, inspired by Catmull’s book on Pixar’s approach to creativity eval.
English
0
0
0
0
Lisa van Gelder
Lisa van Gelder@lisa_van_gelder·
@AnkurHop Yeah it’s tricky. Does thinking about success criteria help? How would you know if an experimental project is going well?
English
1
0
0
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@lisa_van_gelder just saw your talk at CTO Connection’s NY Summit, thanks for some great advice! Did you have any tips on picking the right metrics for R&D teams? DORA doesn’t seem like the right fit for CV/ML research projects.
English
1
0
0
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@lisa_van_gelder Thanks, I think that helps. We do that for applied ML kinds of use cases, but are a little less clear on experimental moonshot type R&D projects.
English
1
0
0
0
Lisa van Gelder
Lisa van Gelder@lisa_van_gelder·
@AnkurHop It’s about figuring what matters to your team. At Avvir our ML team tracks model accuracy & length of feedback loop - ie how long does it take to get an experiment to prod
English
1
0
1
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@mark_riedl @GeorgiaTech @NSF Super interested! As a GT alum, very excited by this work at ML@GT. Will the kickoff talks be recorded? Would love to attend but have a conflict.
English
0
0
0
0
Mark Riedl
Mark Riedl@mark_riedl·
@GeorgiaTech @NSF Research activities will revolve around multi-agent coordination between agents that work on behalf of teams of human caregivers, longitudinal modeling of individuals, speech, dialogue, learning human values, trustworthy AI/ML, explainable AI, and much more.
English
1
0
12
0
Mark Riedl
Mark Riedl@mark_riedl·
Hi folks, @GeorgiaTech is launching one of its new @NSF AI Institutes this Wednesday! See the Institute’s site for kickoff details ai-caring.org The AI-CARING Institute looks at fundamental AI challenges associated with aging at home with dignity
Mark Riedl tweet media
English
1
7
51
0
Ankur Kalra
Ankur Kalra@AnkurHop·
@neilbedwell Love your work (and your space on Edgewood). Minor heads up — the hello bar on your website 404s. Seems like an easy fix in webflow.
English
1
0
0
0
Ankur Kalra retweetledi
ICAD2024
ICAD2024@ICAD_official·
ABBA (Auditory Brown Bag At home) series came back! icad.org/abba-oct-2021/ Oct 27th (Wed) 11 am EDT: Sonification for Process Monitoring: Perspectives and Opportunities in Distributed Systems @AnkurHop, Hop Labs (Industry) and @DrPaulVickers Northumbria University (Academia).
ICAD2024 tweet mediaICAD2024 tweet media
English
0
4
6
0
Ankur Kalra
Ankur Kalra@AnkurHop·
This is why I love @MICCAI_Society. Stumbled into a talk at #MICCAI2021 where Dr Jongho Lee talks about using reinforcement learning (inspired by AlphaGoZero) for RF pulse design to streamline MRI acquisition.
Ankur Kalra tweet media
English
0
0
2
0
Ankur Kalra
Ankur Kalra@AnkurHop·
Really appreciate @mattlungrenMD sharing the honest truth about clinical AI. Six months of data use convos are common, and even then it comes down to who sits on what board, or has a friend who is an exec, or has better mktg. Not perf based or a true marketplace. #MICCAI2021
English
0
0
1
0