Remember kids: Don't burn chrome!
4,595 karma · joined March 8, 2009
https://en.wikipedia.org/wiki/Sausage_Software
Edit: And yes, I saw what you did there.
Location: West coast Canada, remote only.
Technologies: Machine learning/AI, LLM's/RAG, Python, Pytorch/Tensorflow/Keras, A whole host of technologies over 25 year career.
With data science you can translate exactly a great portion of the world into a point on an interconnected web, which then lets you pluck that web to see the reverberations which you can translate back into the world. Without having to pluck the world, which you cannot do.
An ideal fit would be a small/medium sized company with hair on fire green or brownfield problems. I'm currently quite interested in routing/optimization problems, and small AI with logic and external memories.
Currently working on
- ATM skimmer installation identification.
- Elder abuse identification.
- real-time Vehicle routing, Get a couch from where you are, to where you want it to be.
- Failure prediction for a well known stuttgart auto maker.
- Server fleet failure prediction for a well known cloud vendor. Give a 5 minute warning to do load shedding.
- Voice agent in the pre-LLM era. The goal oriented structure used is likely to make a comeback as "router" infra between models.
- Taxi fleet maintenance optimization and failure prediction for a latam taxi company.
- Automated sports highlights - Feed in video, get highlights of diving catches, touchdowns, etc.
- Wound identification and mangement - Identify what sort of wound, label and categorize the tissue.
- Account grading and revenue optimization for ARM.
- Maritime anti-piracy - Identifying pirate skiffs, and at risk vessels.
- Ads or gambling. I'd rather sleep at night. Thanks.Location: West coast Canada, remote only.
Technologies: Machine learning/AI, LLM's/RAG, Python, Pytorch/Tensorflow/Keras, A whole host of technologies over 25 year career.
With data science you can translate exactly a great portion of the world into a point on an interconnected web, which then lets you pluck that web to see the reverberations which you can translate back into the world. Without having to pluck the world, which you cannot do.
An ideal fit would be a small/medium sized company with hair on fire green or brownfield problems. I'm currently quite interested in routing/optimization problems, and small AI with logic and external memories.
Currently working on
- ATM skimmer installation identification.
- Elder abuse identification.
- real-time Vehicle routing, Get a couch from where you are, to where you want it to be.
- Failure prediction for a well known stuttgart auto maker.
- Server fleet failure prediction for a well known cloud vendor. Give a 5 minute warning to do load shedding.
- Voice agent in the pre-LLM era. The goal oriented structure used is likely to make a comeback as "router" infra between models.
- Taxi fleet maintenance optimization and failure prediction for a latam taxi company.
- Automated sports highlights - Feed in video, get highlights of diving catches, touchdowns, etc.
- Wound identification and mangement - Identify what sort of wound, label and categorize the tissue.
- Account grading and revenue optimization for ARM.
- Maritime anti-piracy - Identifying pirate skiffs, and at risk vessels.
- Ads or gambling. I'd rather sleep at night. Thanks.Bloody Commie soda!
This amuses me.
Location: West coast Canada, remote only.
Technologies: Machine learning/AI, LLM's, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
I'm a veteran senior data scientist, who's worked on a number of interesting projects. I'd like to work on more.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. I'm currently quite interested in routing/optimization problems, and AI with logic and external memories.
Some past work:
- real-time Vehicle routing, Get a couch from where you are, to where you want it to be.
- Failiure prediction for a well known stuttgart auto maker.
- Server fleet failure prediction for a well known cloud vendor. Give a 5 minute warning to do load shedding.
- Voice agent in the pre-LLM era. The goal oriented structure used is likely to make a comeback as "router" infra between models.
- Taxi fleet maintenance optimization and failiure prediction for a latam taxi company.
- Automated sports highlights - Feed in video, get highlights of diving catches, touchdowns, etc.
- Wound identification and mangement - Identify what sort of wound, label and categorize the tissue.
- Account grading and revenue optimization for ARM.
- Maritime anti-piracy - Identifying pirate skiffs, and at risk vessels.
- Ads or gambling. I'd rather sleep at night. Thanks.
- Metabolic flux anaylsis - Predicting how metabolic networks consume and produce metabolites.
- AI with logic (Eg: NOT peanuts) and connecting them to interpretable and defensible knowledge bases.Location: West coast Canada, remote only.
Technologies: Machine learning/AI, LLM's, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
I'm a veteran senior data scientist, who's worked on a number of interesting projects. I'd like to work on more.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. I'm currently quite interested in routing/optimization problems, and AI with logic and external memories.
Some past work:
- real-time Vehicle routing, Get a couch from where you are, to where you want it to be.
- Failiure prediction for a well known stuttgart auto maker.
- Server fleet failure prediction for a well known cloud vendor. Give a 5 minute warning to do load shedding.
- Voice agent in the pre-LLM era. The goal oriented structure used is likely to make a comeback as "router" infra between models.
- Taxi fleet maintenance optimization and failiure prediction for a latam taxi company.
- Automated sports highlights - Feed in video, get highlights of diving catches, touchdowns, etc.
- Wound identification and mangement - Identify what sort of wound, label and categorize the tissue.
- Account grading and revenue optimization for ARM.
- Maritime anti-piracy - Identifying pirate skiffs, and at risk vessels.
- Ads or gambling. I'd rather sleep at night. Thanks.
- Metabolic flux anaylsis - Predicting how metabolic networks consume and produce metabolites.
- AI with logic (Eg: NOT peanuts) and connecting them to interpretable and defensible knowledge bases.Well, as I said that's amenable to automated planning.
But what I really, really want is a nice APU with 512GB+ of memory that both the CPU and GPU can access willy nilly.
They had some fancy marketing name for it at the time. But it wasn't on all chips, it should have been. Even if it was dog slow between PCIe GPU and CPU the unified interface would have been the right way to go. Also, amenable to automated scheduling.
The point still stand though, I want entirely unified GPU and CPU memory.
Has this returned? Because for dual gpu/cpu workloads (alpha zero, etc) that would deliver effective “infinite bandwidth” between gpu and cpu. Using an apu of course gets you huge amounts of slowish memory. But being some to fling things around with abandon would be an advantage, particularly for development.
Threads is much worse.
Instead start a spreadsheet with all the people who worked on this product, and contact them as soon as Google announces it is shutting the API down. Or you notice a few people chaining their Linkedin status.
Costs about $800K USD, which is a few tens of millions less than the real thing. Apparently you also see the GTO cars racing in the classic races, as a stand in when the owner owns the real thing.
Location: West coast Canada.
Remote: Yes.
Technologies: Machine learning, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
I'm an experienced Data scientist, who's worked on a bunch of interesting projects. With Covid I've come to realize I need to work in public more often.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. What sort of problems do I like?
I worked with a Stuttgart based auto manufacturer to identify and predict when a part they had would fail. It turns out the answer will be familiar to west coast Canadians - The leaky condo crisis.
I worked with a well known cloud vendor to predict server failiures. Built a few test approaches (DL, anomoly detection, clustering, etc) and went with a simple efficient one that worked well. They got several minutes warning to load shift before a server failed.
A Taxi company in latam hired me to built a fleet maintenance system. It predicted and scheduled preventative maintenance to reduce wear and tear and avoid breakdowns. We also identified the leading cause of wear - leadfoot drivers. So we built a system that would identify aggressive driving.
Another example was a speculative sports highlights annotation system I built with a consultancy. It ingested sports video and identified predetermined highlights in baseball - diving catches in our case.
Who shouldn't contact me.
If your market is Ads or gambling I'm not interested. Sure there's lot of money, but I'd rather sleep at night. Thanks.
What I'm working on right now:
ML for metabolic flux anaylsis - Predicting how metabolic networks consume and produce metabolites. Particulalry useful for potentially starving cancers.
Low compute AI and connecting them to interpretable and defensible knowledge bases. I think these huge models that cost millions to train are the wrong path. I'll show things in public in time.
Location: West coast Canada.
Remote: Yes.
Technologies: Machine learning, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
I'm an experienced Data scientist, who's worked on a bunch of interesting projects. With Covid I've come to realize I need to work in public more often.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. What sort of problems do I like?
I worked with a Stuttgart based auto manufacturer to identify and predict when a part they had would fail. It turns out the answer will be familiar to west coast Canadians - The leaky condo crisis.
I worked with a well known cloud vendor to predict server failiures. Built a few test approaches (DL, anomoly detection, clustering, etc) and went with a simple efficient one that worked well. They got several minutes warning to load shift before a server failed.
A Taxi company in latam hired me to built a fleet maintenance system. It predicted and scheduled preventative maintenance to reduce wear and tear and avoid breakdowns. We also identified the leading cause of wear - leadfoot drivers. So we built a system that would identify aggressive driving.
Another example was a speculative sports highlights annotation system I built with a consultancy. It ingested sports video and identified predetermined highlights in baseball - diving catches in our case.
Who shouldn't contact me.
If your market is Ads or gambling I'm not interested. Sure there's lot of money, but I'd rather sleep at night. Thanks.
What I'm working on right now:
ML for metabolic flux anaylsis - Predicting how metabolic networks consume and produce metabolites. Particulalry useful for potentially starving cancers.
Low compute AI and connecting them to interpretable and defensible knowledge bases. I think these huge models that cost millions to train are the wrong path. I'll show things in public in time.
Location: West coast Canada.
Remote: Yes.
Technologies: Machine learning, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
email in profile.
I'm an experienced Data scientist, who's worked on a bunch of interesting projects. With Covid I've come to realize I need to work in public more often.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. What sort of problems do I like?
I worked with a Stuttgart based auto manufacturer to identify and predict when a part they had would fail. It turns out the answer will be familiar to west coast Canadians - The leaky condo crisis.
I worked with a well known cloud vendor to predict server failures. Built a simple and efficient detection system that worked well. They got several minutes warning to load shift before a server failed.
A Taxi company in latam hired me to built a fleet maintenance system. It predicted and scheduled preventative maintenance to reduce wear and tear and avoid breakdowns. We also identified the leading cause of wear - leadfoot drivers. So we built a system that would identify aggressive driving.
Another example was a speculative sports highlights annotation system I built with a consultancy. It ingested sports video and identified predetermined highlights in baseball - diving catches in our case. The ML pipeline was a CNN model which fed into an RNN. Given a bunch of video clips of diving catches as training data, it could later identify those diving catches and automatically annotate the video feeds.
Who shouldn't contact me.
If your market is Ads or gambling I'm not interested. Sure there's lot of money, but I'd rather sleep at night. Thanks.
What I'm working on right now:
Low compute AI and connecting them to interpretable and defensible knowledge bases. I think these huge models that cost millions to train are the wrong path. I'll show things in public in time.
Location: West coast Canada.
Remote: Yes.
Technologies: Machine learning, Python, Tensorflow/Keras/pytorch, A whole host of technologies over 25 year career.
email in profile.
I'm an experienced Data scientist, who's worked on a bunch of interesting projects. With Covid I've come to realize I need to work in public more often.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly. What sort of problems do I like?
I worked with a Stuttgart based auto manufacturer to identify and predict when a part they had would fail. It turns out the answer will be familiar to west coast Canadians - The leaky condo crisis.
I worked with a well known cloud vendor to predict server failiures. Built a simple and efficient detection system that worked well. They got several minutes warning to load shift before a server failed.
A Taxi company in latam hired me to built a fleet maintenance system. It predicted and scheduled preventative maintenance to reduce wear and tear and avoid breakdowns. We also identified the leading cause of wear - leadfoot drivers. So we built a system that would identify aggressive driving.
Another example was a speculative sports highlights annotation system I built with a consultancy. It ingested sports video and identified predetermined highlights in baseball - diving catches in our case. The ML pipeline was a CNN model which fed into an RNN. Given a bunch of video clips of diving catches as training data, it could later identify those diving catches and automatically annotate the video feeds.
Who shouldn't contact me.
If your market is Ads or gambling I'm not interested. Sure there's lot of money, but I'd rather sleep at night. Thanks.
What I'm working on right now:
Low compute AI and connecting them to interpretable and defensible knowledge bases. I think these huge models that cost millions to train are the wrong path. I'll show things in public in time.
Why did I read this in Connie Sacks' voice talking to George Smiley (The Alec Guinness one)?
I'm sure they'll get Karla (Raghavan), in the end. It's his fanaticism that will do him in.
I missed a few red flags.
The CTO had built a pile of indecipherable mess that runs a good solid business. And that's the rub, it was a good business. The CTO had build a steaming pile, but had it all in his head. If others couldn't wade through it, well, it was a good business.
Being a good business, they could afford to run through bodies. I discovered my "coworkers" were all contract. Another red flag.
So as my work extended out into the rest of the system I slowed down, and was eventually moved on. After all the CTO who wrote most of it could handle it, why couldn't I?
I'll take the Lisp over 6502 assembly any day of the week.
Ask me how I know.
Or even worse, that pile of indecipherable mess runs a good solid business.
Ask me how I know.
Edit: They were CTO, not CEO.
Can we just remove Nigeria from the internet?
Which builds up from first principles a ballistic missile defence system. Always useful to have.
Tactical and Strategic Missile Guidance, Seventh Edition https://www.amazon.com/Tactical-Strategic-Missile-Guidance-S...
Tactical missile warheads https://www.amazon.com/Tactical-Warheads-Progress-Astronauti...
Location: West coast Canada.
Remote: Yes.
Technologies: Machine learning, Python, Tensorflow/Keras/pytorch, spark, Scala, AWS, kafka, HDFS, etc. A whole host of technologies over 20 year career.
email in profile.
I'm an experienced Data scientist, who's worked on a bunch of interesting projects. With Covid I've come to realize I need to work in public more often.
So an ideal fit would be a small/medium company with hair on fire problems that need solving that I can talk about publicly.
What sort of problems do I like?
A Taxi company in latam hired me to built a fleet maintenance system. It predicted and scheduled preventative maintenance to reduce wear and tear and avoid breakdowns. We also identified the leading cause of wear - leadfoot drivers. So we built a system that would identify aggressive driving.
Another example was a speculative sports highlights annotation system I built with a consultancy. It ingested sports video and identified predetermined highlights in baseball - diving catches in our case. The ML pipeline was a CNN model which fed into an RNN. Given a bunch of video clips of diving catches as training data, it could later identify those diving catches and automatically annotate the video feeds.
Who shouldn't contact me.
If your market is Ads or gambling I'm not interested. Sure there's lot of money, but I'd rather sleep at night. Thanks.