Hacker News new | past | comments | ask | show | jobs | submit | cstavish's comments login

Proscia | Front-end Engineer, Back-end Engineer, DevOps, Research Engineer, Machine Learning Systems Engineer | Philadelphia, PA | Full-time | Onsite

The data to fight cancer is in images--specifically, gigapixel scans of tissue. Proscia is working to make pathology diagnosis more accurate and more efficient through the thoughtful application of artificial intelligence. Proscia's software is in use in laboratories today and has been adopted by world-renowned medical centers.

Our data currency is high-resolution virtual microscopy images, affording unique challenges and growth potential to all technical roles. This isn't another digital health CRUD app. If you're interested working on a meaningful problem alongside talented and good-natured colleagues, send me an email at coleman [at] proscia.com.

We offer competitive salaries, equity, health/dental/vision/disability and relocation assistance.


Those familiar with the field will recognize an augmented reality microscope as an incomplete alternative to, or stepping-stone towards AI applied to whole-slide images, which are multi-resolution images of more or less identical quality quality to that of a microscope.

For some use-cases, deploying AI to microscopic fields of view is a viable, lower-cost alternative to creating whole-slide images and running AI on them in their entirety (whole-slide scanners are a bit expensive). Specifically, if a pathologists identifies a suspicious region, the augmented scope can provide useful support. However, many types of anatomic pathology assessment require laborious review of several slides. Only AI applied to whole-slide images can pre-identify rare events or "hot spots", saving pathologist time while improving diagnostic confidence.


Exhaustive search is definitely a high value option. But let's consider dermpath. Where the entire lesion is probably in the field of view at 4x, but the number of possible classes may be high (bullous lesions, anyone?). In that case, "90% Bowen's disease, 3% inflamed Seborhheic keratosis, and 7% other" is useful, especially if you knew the machine had been trained on everything in Weedon.


I agree with you, and there is several companies already doing this for whole slides. In addition, people should also understand that more than 1 slide is necessary for a high quality analysis of the cancer. From the top of my head, I remember Definiens (http://definiens.com) that develops software that does whole slide analysis and HalioDX (http://www.haliodx.com/) that does the same and use it as auxiliary for colon cancer treatment.


> (whole-slide scanners are a bit expensive)

Huh... Why? It _seems_ like a problem with an easy (and cheap) solution, given how readily available (and cheap!) suitably precise cartesian bots are these days, combined with high-quality digital microscopes and the state of modern image alignment algorithms.

What am I missing here?


A lot of that cost comes from all the supporting hardware and certified software. Plus a lot of the machines have to be designed for high throughput applications.


I work a bit in the field and I think it's a completely impractical but fun hack.


I work in the field and am about to be sent to a remote location for 2 years. Literally the only pathologist for 100s of miles. This would be more than a little nice to have.


Sapiens?


Sapiens: A Brief History of Humankind

a sort of maco-history of human civilization. Highly recommend

https://en.wikipedia.org/wiki/Sapiens%3A_A_Brief_History_of_...


Ha, yes, sorry. Went up a level too high.


In a similar spirit, the Bloomberg Businessweek magazine used to have zany cover designs (my favorite featured a sweaty Steve Ballmer in a Clippers jersey dribbling a basketball captioned "BASKET BALLMER") but they've since switched to plain, solid colors.


Yeah, my favorite was the animated GIF of a flying Jeff Bezos inside a cardboard box.


Programming Pearls by Jon Bentley.


Accredited investors/VC (disclaimer: I am neither, but co-founded a VC-backed startup) are often wary of crowd funding because of the idea that with more investors, the more headaches that can occur with additional capital raises. Having a big cap table doesn't make company operations easier. I honestly don't know how valid those viewpoints are, but that's what I've heard.


I believe it. I also know that Realtors aren't too fond of For Sale By Owner houses. I suspect this comes down to a similar effect.


> I also know that Realtors aren't too fond of For Sale By Owner houses

This scares agents because a wide-spread DIY real estate industry kills all of the middlemen (themselves)


This is an interesting conversation to have here really.

'Never ask a barber if he thinks you need a haircut'


Yeah, it's kind of 'meta'.

Though the interesting thing is, Y Combinator is interested in basic income. That's effectively like crowd-seeding, which makes all the 'cut out the middleman' stuff more practical.

If you're expecting 'disruption' and embracing it, and you see disruption coming for your own livelihood, the sophisticated response is to try and find a place for yourself in the new situation, rather than stop things from changing at all. In that light, Y Combinator looking at ways to cut out SV incubators could be prescient.

It's like with me: I run a very functional small recording studio as part of my business, and it's well on the way to handling the live recording of traditional bands and their instrumentation. Yet I'm a lot more interested in finding unusual ways to support electronic genres (a strikingly different skillset).


It's funny here too, because I would assume that crowd funding would put you on a more conservative path. I like the Idea of taking $100k seed money an building a company that turns a profit. Not one that IPO's and makes us fuck you money, just one that pays back my investors at better than market, and pays my employees.

But you're signing up for a lot of work, not a sprint, and that's kinda the antithesis to 'changing the world' without making a profit then going to IPO.

I don't want to gamble on something, I want to build something with a less sophisticated valuation model where the revenue plan is clear from the start.


I don't see why this is an issue though, just promise quarterly updates and clearly state rights on future raises for existing investors. They can follow-on or not. All of these issues can be solved, it's just more excuses. It should be easier to have 100 investors who all agree on everything than 4 investors you have to baby and cajole and manage their whining because they feel entitled to personal service.


That assumes that the company proceeds through successive rounds with an upwards trajectory. If the company falters and takes a down round the initial investors may disagree with the direction or even sue.

If the company needs to be recapitalized or restructured under a time crunch, tracking down 100 loosely involved investors and getting them to sign the deal may not be feasible.


Probably would have been better to have arbitration instead of lawsuits written into the law. (to make it work not for investors)


The way AngelList (and I think FundersClub, too) handle it is to set up an LLC. The investors are then partners in an LLC, not direct shareholders.


> The investors are then partners in an LLC

Pedantically, "members" not "partners"; an LLC is not a partnership.


Seedrs also does it like this.


> as I don't know if pulling venture capital is as easy as selling the stock.

There's generally minimal liquidity. During a round, an existing investor may have the opportunity to sell some shares to new/other investors, but if she knows something that's not coming out during diligence, there's definitely something fishy going on. If the round is shaping up to be a major up-round, maybe an early investor wants to lock in a good return, but that's beside the point here. And then of course if things really aren't going well for the company, you're looking at the bad kind of liquidity event--a liquidation.

So if a VC wants to pull out based on a negative hunch, it's probably either impossible or the signal itself will doom the company if it wasn't already doomed.

Just my 2 cents as a first-time founder.


Sounds about right to me.


Thanks a million, from me and the people who invested in my startup.

edit: apparently SNI isn't supported on legacy dynos.


In 2014, while in college, I did a co-op rotation at the Bombardier Transportation facility in Pittsburgh, PA. The rail division may be based in Germany, but there's a sizable rail engineering contingent in Pittsburgh. I worked with a lot of good people and engineers there--I hope they do alright. Even in 2014 there was grumbling by the engineers about upper management, and I remember attending an "all-hands" meeting where a visiting executive talked of belt-tightening. In a fairly tone-deaf delivery, he spoke of limiting his own business-class international travel.


>In a fairly tone-deaf delivery, he spoke of limiting his own business-class international travel.

I am reliably informed that business travel sucks donkey ass, (almost) regardless of class. It's tone deaf because travel has positive connotations to most people, not because he's out of touch with reality. To a business traveler, economy class is adding insult to injury.


I am reliably informed (by Kayak) that the cheapest direct flight from JFK to Heathrow, one week from today, costs $901 in economy class on British Airways, while the cheapest business class ticket costs $5451 on Virgin Atlantic.

The saying "A fool and his money are soon parted" extends to corporations, too.


Yes and no. I remember back in the day an exec taking a Concorde flight from NYC to London to tell us that if we could all manage to save $5/day, there would be no layoffs... Yeah, he was wrong about that, but he would have gotten loads of airmiles...


I'm the same with my interactions with Bombardier. Every time I've come into contact with them in business, or through networking functions, I've noticed what a good group of people they've been.


I use JSONB columns for similar use-cases, but to play devil's advocate, you can accomplish that a fourth way, which is almost certainly better than 1 or 2. A table for fields, one row per field. A table for forms, which has a many-to-many relation to fields. Entries in a link table compose a form of arbitrary fields. Answers can be stored in a separate responses table, indexed by form_id and column_id. I don't know enough about database implementation to speculate on how that would perform at scale, but conceptually that's how I think of the problem.


This is roughly EAV. EAV is where you have a with a schema:

    entity_id | attribute_id | value
EAV is typically considered to be an anti-pattern for several reasons: it becomes very expensive to rematerialize a entity with all of its attributes/values, it becomes difficult to have any kind of constraints on the values (this is also a problem with jsonb), and it's hard for the database to maintain good statistics on the values of different attributes as most databases don't keep multi-column statistics. Don't worry, I've had similar ideas before.


EAV is an anti-pattern?

If you don't have JSONB, EAV is the only remotely-reasonable way to implement user-defined fields (e.g. product-specific attributes in e-commerce).


If you want to allow user-defined fields in a relational database, your realistic are either EAV or stuff json into a text column. EAV, if done extremely carefully, can be a good solution, but 99% of the time, it's going to be a huge pain.


> stuff json into a text column

You did see the article was about JSONB, which is significantly more sophisticated than "json in a text column", yes?


Yes, the person I was replying to was asking about what to do if you don't have jsonb.


Sounds kind of like the EAV relational data model: https://en.wikipedia.org/wiki/Entity%E2%80%93attribute%E2%80....


I would expect that to perform poorly at scale.

There is also a fifth option, though it is not very space efficient. That is to store every field in the JSONB column. Personally, that is what I would do.

This approach is known to be used successfully by many high scale companies. It ensures the highest degree of flexibility and still allows for full indexing of the fields. The schema would need to be enforced by your application, but that should happen anyway.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: