Category projects

Making this blog

This is an old post from a now-deprecated website. I’m keeping it around only for historical preservation reasons. Not much of it is relevant to the website you’re currently reading on.

Category engineering

Consistency is better than surge; or how to not burn out

I spent way too much time and effort working on this blog. I forgot consistency mattered over immediate output. I was burnt out. I put a too much effort writing and coding this website that it stopped being fun or interesting. Taking things easy is the way to go. Consistency...

Bringing lessons from Japanese temple constructions to software maintenance

Don’t entomb your applications. Make them flexible, be ready to make gradual changes. That way, they won’t need an overhaul from scratch when the time comes.

Diverse interests help us understand the world better

Having a diverse set of interests allows you to look at the world in different ways. Pick up a weird hobby, or a new craft. If nothing else, it’ll help you do your job better.

The duct-tape and strings approach to software is unfairly maligned

The “duct-tape and strings” approach to building software prioritizes functionality and rapid iteration. That comes at the cost of design perfection and technical debt. It raises eyebrows among proponents of careful planning. However, it’s a valid approach to real-world pressures and can lead to successful results.

How to prioritize profit while supporting innovation

It’s easy to be seduced by the latest ‘innovation’ fads. But it’s important to focus on aligning innovation with core business goals and customer needs. Doing so will protect teams from the innovation trap and achieve long-term success. A lot of this essay is inspired by Christen’s book ‘The Innovation...

Creating and leading high-trust engineering teams for success

Your organization needs ‘high-trust’ engineering teams. A strong team with high intra-team trust, and good relations between leadership and the team, is a powerful and effective institution. Such high-trust engineering teams are a force to be reckoned with. Here’s how to create such teams, how you’ll benefit from doing so,...

How to evaluate the right choice of technology for your team

Choosing the right tool or library for software projects can feel like navigating a minefield. This essay will offer a practical framework for evaluating competing technologies. You’ll be able to make better and informed choices that align with your project’s needs. There is no ‘perfect option’, so one must understand...

How my Fermentation Hobby Helps Me as a Software Engineer

For a software engineer, technical expertise is clearly of the highest importance. However, how they approach solving problems can significantly shape how they get their everyday tasks accomplished too. Through a hobby of mine I’ve discovered an unexpected wellspring of professional growth. Fermentation. This seemingly random hobby has become a...

What's next for AI? My predictions for the next 10 years

Explainable AI, optical AI, analog AI, and meta-learning. I predict the culmination of optimizing massive generative models will result generative AI being turned into a commodity, and these four areas will be the next frontiers of innovation in machine learning.

Book Review: Learning to Learn and the Navigation of Moods by Gloria P. Flores

One can create a mindset highly open and receptive to learning quickly by managing one’s moods. That’s what learning researcher and psychologist Gloria Flores argues in her book Learning to Learn. She argues that only by modifying the less-productive (for learning) moods towards more productive ones, does learning get easier....

Understanding what kind of leader you are: a new lens to look at Engineering leadership

Here’s my approach to to help engineering leaders understand themselves. It will allow them leverage their skillsets the best.

How our squad integrated data scientists and software engineers

Our team integrated data scientists and engineers into a single team. We learned important lessons as we tested different team structures and planning regimen. This essay takes us through our journey.

A tale of the treacherous task of Kubernetes upgrade

This is a tale of unmitigated blast radius in networking. I broke two of our developer environments the second day of joining the infrastructure squad. We discovered several potential issues with our workflow and engineering practices. Let me explain how that happened, and what we did to fix it. We’ll...

Infrastructure management: from chaos to Cloudformation CDK to confusion

Our team decided to use Amazon’s infrastructure-management tooling versus other available tooling a while back. We are now reconsidering the decision.

Don't hoard your engineering players, let them out in the field

Well-run organizations can fall into the trap of ‘hoarding’ their star performers. They keep them ‘on the bench’, not deploying them for big, risky, challenging projects. That’s an unnecessary inefficiency. As in sports, allowing your strongest players to go out to the field gives them practice. It also improves general...

Category software

Bringing lessons from Japanese temple constructions to software maintenance

Don’t entomb your applications. Make them flexible, be ready to make gradual changes. That way, they won’t need an overhaul from scratch when the time comes.

Diverse interests help us understand the world better

Having a diverse set of interests allows you to look at the world in different ways. Pick up a weird hobby, or a new craft. If nothing else, it’ll help you do your job better.

A tale of the treacherous task of Kubernetes upgrade

This is a tale of unmitigated blast radius in networking. I broke two of our developer environments the second day of joining the infrastructure squad. We discovered several potential issues with our workflow and engineering practices. Let me explain how that happened, and what we did to fix it. We’ll...

Infrastructure management: from chaos to Cloudformation CDK to confusion

Our team decided to use Amazon’s infrastructure-management tooling versus other available tooling a while back. We are now reconsidering the decision.

Don't hoard your engineering players, let them out in the field

Well-run organizations can fall into the trap of ‘hoarding’ their star performers. They keep them ‘on the bench’, not deploying them for big, risky, challenging projects. That’s an unnecessary inefficiency. As in sports, allowing your strongest players to go out to the field gives them practice. It also improves general...

Category leadership

The duct-tape and strings approach to software is unfairly maligned

The “duct-tape and strings” approach to building software prioritizes functionality and rapid iteration. That comes at the cost of design perfection and technical debt. It raises eyebrows among proponents of careful planning. However, it’s a valid approach to real-world pressures and can lead to successful results.

Srini Pillay's Tinker Dabble Try: The Power of an Unfocused Mind, a book review

Planned “unfocused” activities can enhance your cognitive toolkit. So says Dr Srini Pillay In “Tinker Dabble Doodle Try: The Power of an Unfocused Mind”. In the book Dr. Pillay challenges the idea that laser-sharp focus is the key to success.

How to prioritize profit while supporting innovation

It’s easy to be seduced by the latest ‘innovation’ fads. But it’s important to focus on aligning innovation with core business goals and customer needs. Doing so will protect teams from the innovation trap and achieve long-term success. A lot of this essay is inspired by Christen’s book ‘The Innovation...

Creating and leading high-trust engineering teams for success

Your organization needs ‘high-trust’ engineering teams. A strong team with high intra-team trust, and good relations between leadership and the team, is a powerful and effective institution. Such high-trust engineering teams are a force to be reckoned with. Here’s how to create such teams, how you’ll benefit from doing so,...

The Checklist Manifesto: A Guide to Better Software Engineering

A review of Atul Gawande’s seminal book ~ The Checklist Manifesto ~ and the lessons it holds for engineers and engineering leaders.

A blogger's plea (and a proposal) to Google: bring back Blogger!

This essay proposes a plan for revitalization of Blogger through a premium subscription model, cool exciting features powered by Google’s AI capabilities, and a renewed focus on user needs. Blogger, once a dominant force in the blogosphere, has stagnated in recent years. While competitors like Medium and substack have thrived...

Book review of David Epstein’s Range: Why Generalists Triumph in a Specialized World

Author David Epstein argues that individuals who embrace diverse experiences and develop a broader range of skills outperform specialists in complex and unpredictable environments. The conclusion flies in the face of the conventional wisdom that values early specialization and deliberate practice in a single domain. Epstein’s central thesis in his...

Book Review: Learning to Learn and the Navigation of Moods by Gloria P. Flores

One can create a mindset highly open and receptive to learning quickly by managing one’s moods. That’s what learning researcher and psychologist Gloria Flores argues in her book Learning to Learn. She argues that only by modifying the less-productive (for learning) moods towards more productive ones, does learning get easier....

Understanding what kind of leader you are: a new lens to look at Engineering leadership

Here’s my approach to to help engineering leaders understand themselves. It will allow them leverage their skillsets the best.

How our squad integrated data scientists and software engineers

Our team integrated data scientists and engineers into a single team. We learned important lessons as we tested different team structures and planning regimen. This essay takes us through our journey.

Don't hoard your engineering players, let them out in the field

Well-run organizations can fall into the trap of ‘hoarding’ their star performers. They keep them ‘on the bench’, not deploying them for big, risky, challenging projects. That’s an unnecessary inefficiency. As in sports, allowing your strongest players to go out to the field gives them practice. It also improves general...

Category cloudformation

A tale of the treacherous task of Kubernetes upgrade

This is a tale of unmitigated blast radius in networking. I broke two of our developer environments the second day of joining the infrastructure squad. We discovered several potential issues with our workflow and engineering practices. Let me explain how that happened, and what we did to fix it. We’ll...

Infrastructure management: from chaos to Cloudformation CDK to confusion

Our team decided to use Amazon’s infrastructure-management tooling versus other available tooling a while back. We are now reconsidering the decision.

Category terraform

Infrastructure management: from chaos to Cloudformation CDK to confusion

Our team decided to use Amazon’s infrastructure-management tooling versus other available tooling a while back. We are now reconsidering the decision.

Category infrastructure

The story of what it took to setup NVIDIA GPU drivers and time-slicing in our GPU EKS cluster

I spent six months working on a work project that I thought would take two weeks. This essay narrates the adventure in implementing GPU time slicing on our EKS kubernetes cluster. It began as a seemingly straightforward task – installing the gpu-operator. The work morphed into a long-lasting exploration that...

Software infrastructure and databases workshop for interns and new engineers

This is the syllabus I originally designed to train Hack.diversity fellows for a 8-week workship (1 hour weekly classes, 2 hours of expected assignment load) to help them perform well on their internship/co-op interviews. I’ve worked on this with the engineer I mentor with the organization, and will be proposing...

A tale of the treacherous task of Kubernetes upgrade

This is a tale of unmitigated blast radius in networking. I broke two of our developer environments the second day of joining the infrastructure squad. We discovered several potential issues with our workflow and engineering practices. Let me explain how that happened, and what we did to fix it. We’ll...

Category kubernetes

The story of what it took to setup NVIDIA GPU drivers and time-slicing in our GPU EKS cluster

I spent six months working on a work project that I thought would take two weeks. This essay narrates the adventure in implementing GPU time slicing on our EKS kubernetes cluster. It began as a seemingly straightforward task – installing the gpu-operator. The work morphed into a long-lasting exploration that...

A tale of the treacherous task of Kubernetes upgrade

This is a tale of unmitigated blast radius in networking. I broke two of our developer environments the second day of joining the infrastructure squad. We discovered several potential issues with our workflow and engineering practices. Let me explain how that happened, and what we did to fix it. We’ll...

Category teamwork

How our squad integrated data scientists and software engineers

Our team integrated data scientists and engineers into a single team. We learned important lessons as we tested different team structures and planning regimen. This essay takes us through our journey.

Category styles

Understanding what kind of leader you are: a new lens to look at Engineering leadership

Here’s my approach to to help engineering leaders understand themselves. It will allow them leverage their skillsets the best.

Category learning

Book Review: Learning to Learn and the Navigation of Moods by Gloria P. Flores

One can create a mindset highly open and receptive to learning quickly by managing one’s moods. That’s what learning researcher and psychologist Gloria Flores argues in her book Learning to Learn. She argues that only by modifying the less-productive (for learning) moods towards more productive ones, does learning get easier....

Category skill

Book Review: Learning to Learn and the Navigation of Moods by Gloria P. Flores

One can create a mindset highly open and receptive to learning quickly by managing one’s moods. That’s what learning researcher and psychologist Gloria Flores argues in her book Learning to Learn. She argues that only by modifying the less-productive (for learning) moods towards more productive ones, does learning get easier....

Category book-review

Srini Pillay's Tinker Dabble Try: The Power of an Unfocused Mind, a book review

Planned “unfocused” activities can enhance your cognitive toolkit. So says Dr Srini Pillay In “Tinker Dabble Doodle Try: The Power of an Unfocused Mind”. In the book Dr. Pillay challenges the idea that laser-sharp focus is the key to success.

The Checklist Manifesto: A Guide to Better Software Engineering

A review of Atul Gawande’s seminal book ~ The Checklist Manifesto ~ and the lessons it holds for engineers and engineering leaders.

Book review of David Epstein’s Range: Why Generalists Triumph in a Specialized World

Author David Epstein argues that individuals who embrace diverse experiences and develop a broader range of skills outperform specialists in complex and unpredictable environments. The conclusion flies in the face of the conventional wisdom that values early specialization and deliberate practice in a single domain. Epstein’s central thesis in his...

Book Review: Learning to Learn and the Navigation of Moods by Gloria P. Flores

One can create a mindset highly open and receptive to learning quickly by managing one’s moods. That’s what learning researcher and psychologist Gloria Flores argues in her book Learning to Learn. She argues that only by modifying the less-productive (for learning) moods towards more productive ones, does learning get easier....

Category ai

Running large language models (LLM) locally on your Mac using OLLAMA

This is an introduction to Large Language Models (LLMs) and an instruction to running them locally on a Mac using OLLAMA.

LLM's will be the next spellcheck assistant, not the next robotic overlords!

I’ve changed my opinion on genAI and LLM’s a few times now. As technology, politics, and economics around these technologies change, my beliefs have evolved. In this essay I propose that generative AI will be mostly a collaborative revolution. Generative models can magnify human abilities and make collaboration easier. We...

What's next for AI? My predictions for the next 10 years

Explainable AI, optical AI, analog AI, and meta-learning. I predict the culmination of optimizing massive generative models will result generative AI being turned into a commodity, and these four areas will be the next frontiers of innovation in machine learning.

The risk of production, customer-facing LLM's let lose

Readers of technology-adjacent news might remember the recent case where a Canadian passenger, Jake Moffatt, sued Air Canada after its online chatbot misinformed him about bereavement fares, costing him hundreds of dollars? Air Canada ended up losing the case, and now must stand by the commitment made by its AI...

The tasks LLM's are good at, and the tasks they're sh**t at: a personal review

My views on how LLM’s might impact our economies and personal lives have shifted a lot in recent years. This is an explanation of that shift. LLM’s can be good at some tasks, and awful at others. Being able to understand the distinction between the two modes helped make my...

Category llm

Running large language models (LLM) locally on your Mac using OLLAMA

This is an introduction to Large Language Models (LLMs) and an instruction to running them locally on a Mac using OLLAMA.

Software infrastructure and databases workshop for interns and new engineers

This is the syllabus I originally designed to train Hack.diversity fellows for a 8-week workship (1 hour weekly classes, 2 hours of expected assignment load) to help them perform well on their internship/co-op interviews. I’ve worked on this with the engineer I mentor with the organization, and will be proposing...

LLM's will be the next spellcheck assistant, not the next robotic overlords!

I’ve changed my opinion on genAI and LLM’s a few times now. As technology, politics, and economics around these technologies change, my beliefs have evolved. In this essay I propose that generative AI will be mostly a collaborative revolution. Generative models can magnify human abilities and make collaboration easier. We...

What's next for AI? My predictions for the next 10 years

Explainable AI, optical AI, analog AI, and meta-learning. I predict the culmination of optimizing massive generative models will result generative AI being turned into a commodity, and these four areas will be the next frontiers of innovation in machine learning.

The risk of production, customer-facing LLM's let lose

Readers of technology-adjacent news might remember the recent case where a Canadian passenger, Jake Moffatt, sued Air Canada after its online chatbot misinformed him about bereavement fares, costing him hundreds of dollars? Air Canada ended up losing the case, and now must stand by the commitment made by its AI...

The tasks LLM's are good at, and the tasks they're sh**t at: a personal review

My views on how LLM’s might impact our economies and personal lives have shifted a lot in recent years. This is an explanation of that shift. LLM’s can be good at some tasks, and awful at others. Being able to understand the distinction between the two modes helped make my...

Category collaboration

Software infrastructure and databases workshop for interns and new engineers

This is the syllabus I originally designed to train Hack.diversity fellows for a 8-week workship (1 hour weekly classes, 2 hours of expected assignment load) to help them perform well on their internship/co-op interviews. I’ve worked on this with the engineer I mentor with the organization, and will be proposing...

How to evaluate the right choice of technology for your team

Choosing the right tool or library for software projects can feel like navigating a minefield. This essay will offer a practical framework for evaluating competing technologies. You’ll be able to make better and informed choices that align with your project’s needs. There is no ‘perfect option’, so one must understand...

LLM's will be the next spellcheck assistant, not the next robotic overlords!

I’ve changed my opinion on genAI and LLM’s a few times now. As technology, politics, and economics around these technologies change, my beliefs have evolved. In this essay I propose that generative AI will be mostly a collaborative revolution. Generative models can magnify human abilities and make collaboration easier. We...

The risk of production, customer-facing LLM's let lose

Readers of technology-adjacent news might remember the recent case where a Canadian passenger, Jake Moffatt, sued Air Canada after its online chatbot misinformed him about bereavement fares, costing him hundreds of dollars? Air Canada ended up losing the case, and now must stand by the commitment made by its AI...

The tasks LLM's are good at, and the tasks they're sh**t at: a personal review

My views on how LLM’s might impact our economies and personal lives have shifted a lot in recent years. This is an explanation of that shift. LLM’s can be good at some tasks, and awful at others. Being able to understand the distinction between the two modes helped make my...

Category blogger

A blogger's plea (and a proposal) to Google: bring back Blogger!

This essay proposes a plan for revitalization of Blogger through a premium subscription model, cool exciting features powered by Google’s AI capabilities, and a renewed focus on user needs. Blogger, once a dominant force in the blogosphere, has stagnated in recent years. While competitors like Medium and substack have thrived...

Category product

A blogger's plea (and a proposal) to Google: bring back Blogger!

This essay proposes a plan for revitalization of Blogger through a premium subscription model, cool exciting features powered by Google’s AI capabilities, and a renewed focus on user needs. Blogger, once a dominant force in the blogosphere, has stagnated in recent years. While competitors like Medium and substack have thrived...

Category review

The Checklist Manifesto: A Guide to Better Software Engineering

A review of Atul Gawande’s seminal book ~ The Checklist Manifesto ~ and the lessons it holds for engineers and engineering leaders.

Category fermentation

How my Fermentation Hobby Helps Me as a Software Engineer

For a software engineer, technical expertise is clearly of the highest importance. However, how they approach solving problems can significantly shape how they get their everyday tasks accomplished too. Through a hobby of mine I’ve discovered an unexpected wellspring of professional growth. Fermentation. This seemingly random hobby has become a...

Category hobbies

Diverse interests help us understand the world better

Having a diverse set of interests allows you to look at the world in different ways. Pick up a weird hobby, or a new craft. If nothing else, it’ll help you do your job better.

How my Fermentation Hobby Helps Me as a Software Engineer

For a software engineer, technical expertise is clearly of the highest importance. However, how they approach solving problems can significantly shape how they get their everyday tasks accomplished too. Through a hobby of mine I’ve discovered an unexpected wellspring of professional growth. Fermentation. This seemingly random hobby has become a...

Category tooling

How to evaluate the right choice of technology for your team

Choosing the right tool or library for software projects can feel like navigating a minefield. This essay will offer a practical framework for evaluating competing technologies. You’ll be able to make better and informed choices that align with your project’s needs. There is no ‘perfect option’, so one must understand...

Category architecture

The duct-tape and strings approach to software is unfairly maligned

The “duct-tape and strings” approach to building software prioritizes functionality and rapid iteration. That comes at the cost of design perfection and technical debt. It raises eyebrows among proponents of careful planning. However, it’s a valid approach to real-world pressures and can lead to successful results.

How to prioritize profit while supporting innovation

It’s easy to be seduced by the latest ‘innovation’ fads. But it’s important to focus on aligning innovation with core business goals and customer needs. Doing so will protect teams from the innovation trap and achieve long-term success. A lot of this essay is inspired by Christen’s book ‘The Innovation...

Creating and leading high-trust engineering teams for success

Your organization needs ‘high-trust’ engineering teams. A strong team with high intra-team trust, and good relations between leadership and the team, is a powerful and effective institution. Such high-trust engineering teams are a force to be reckoned with. Here’s how to create such teams, how you’ll benefit from doing so,...

Category trust

Creating and leading high-trust engineering teams for success

Your organization needs ‘high-trust’ engineering teams. A strong team with high intra-team trust, and good relations between leadership and the team, is a powerful and effective institution. Such high-trust engineering teams are a force to be reckoned with. Here’s how to create such teams, how you’ll benefit from doing so,...

Category workshop

Software infrastructure and databases workshop for interns and new engineers

This is the syllabus I originally designed to train Hack.diversity fellows for a 8-week workship (1 hour weekly classes, 2 hours of expected assignment load) to help them perform well on their internship/co-op interviews. I’ve worked on this with the engineer I mentor with the organization, and will be proposing...

Category innovation

Do good things, don't worry about scale, don't let systemic issues stop you

This is a lesson that took me a decade to fully ‘get’. You don’t have to worry about the systematic issues to bring change. You can make changes as an individual at a small scale.

How to prioritize profit while supporting innovation

It’s easy to be seduced by the latest ‘innovation’ fads. But it’s important to focus on aligning innovation with core business goals and customer needs. Doing so will protect teams from the innovation trap and achieve long-term success. A lot of this essay is inspired by Christen’s book ‘The Innovation...

Category data-science

The story of what it took to setup NVIDIA GPU drivers and time-slicing in our GPU EKS cluster

I spent six months working on a work project that I thought would take two weeks. This essay narrates the adventure in implementing GPU time slicing on our EKS kubernetes cluster. It began as a seemingly straightforward task – installing the gpu-operator. The work morphed into a long-lasting exploration that...

Category tutorial

Running large language models (LLM) locally on your Mac using OLLAMA

This is an introduction to Large Language Models (LLMs) and an instruction to running them locally on a Mac using OLLAMA.

Category creativity

Srini Pillay's Tinker Dabble Try: The Power of an Unfocused Mind, a book review

Planned “unfocused” activities can enhance your cognitive toolkit. So says Dr Srini Pillay In “Tinker Dabble Doodle Try: The Power of an Unfocused Mind”. In the book Dr. Pillay challenges the idea that laser-sharp focus is the key to success.

Category rewrite

Bringing lessons from Japanese temple constructions to software maintenance

Don’t entomb your applications. Make them flexible, be ready to make gradual changes. That way, they won’t need an overhaul from scratch when the time comes.

Category blogging

Consistency is better than surge; or how to not burn out

I spent way too much time and effort working on this blog. I forgot consistency mattered over immediate output. I was burnt out. I put a too much effort writing and coding this website that it stopped being fun or interesting. Taking things easy is the way to go. Consistency...

Category burnout

Consistency is better than surge; or how to not burn out

I spent way too much time and effort working on this blog. I forgot consistency mattered over immediate output. I was burnt out. I put a too much effort writing and coding this website that it stopped being fun or interesting. Taking things easy is the way to go. Consistency...

Category inspiration

Do good things, don't worry about scale, don't let systemic issues stop you

This is a lesson that took me a decade to fully ‘get’. You don’t have to worry about the systematic issues to bring change. You can make changes as an individual at a small scale.