How Generalist Approach Can Help Me Beat Hyper Specialists in VLSI.
Today I want to discuss on rather very unusual Topic of Generalist vs Specialist in VLSI, based on a book I read recently named “Range” from David Epstein. Based on the limited knowledge I have I would like to break few stereotypes and demystify a lot of “myths” which has been inculcated in most of our Brains by people around us. The article aims to encourage people to develop a wide range of skills rather than confining themselves to a single particular area of interest.
In this article, if I am referring to a Generalist means a Person with a “T” based skill set, Horizontal line represent the breadth of skills in his Life ( Also called “Range”) and Vertical Line refers to the Depth of a Single Skill, whereas a Specialist refers to someone who has no idea of other domains and spends most of his time digging into a specialized area of his/her interest.
To summarize my career in VLSI below are a few of the harsh realities which I faced once I entered the Chip Designing World.
- An RTL Design engineer thinking Backend Work is all about Setup/Hold violations and not many challenges.
- A Place&Route (P&R)/STA/DFT Engineer thinking RTL code stays same and very less scope of designing a new module or Feature in the ASIC.
- An STA/DFT engineer thinking P&R work is boring and vice versa.
- Both STA/DFT and P&R engineers thinking Physical verification/Analog layout engineers do not exist on the Planet and thinking Physical Verification or Analog Layout work is for Diploma Graduates only. etc
I don’t want to Blame anyone Single Person, Together we all are responsible for this kind of disaster which has happened and continues to happen in the future as well. The main reason being, all the above-mentioned engineers are mostly Specialists. Had there been more Generalists they definitely would have given better respect for people working in other domains because of the inter-domain knowledge they have in other fields.
Most of the experienced people in VLSI get worse with experience due to their narrow-mindedness. More experience breeds confidence but not skills — a dangerous combination.
Experience is like a double-edged sword, if used properly can defeat the competition, else it can lead to a Tragic Career in the Long run!
Let’s Dig into more details where I can share few real-life scenarios wherein case if I had known other domain knowledge could have saved a lot of Money and Time for my Clients/Employers.
The Cult of Head Start
Most of us feel that early specialization and Good Head Start is what matters, to be ultimately successful in VLSI. I see people who are beginning their careers in one domain are too hesitant to try out other domain tasks or at least read and gain some knowledge on other Disciplines. The more you sample a wide variety of tasks your seniority will have some value and purpose and you can enjoy a fruitful career in the long run. In my case I started as a P&R engineer later I moved to Full Chip STA, and now I am working on Synthesis/Formality. If I am given the chance I want to explore CDC/DFT in more detail in the future. I don’t care whether I am going deeper in the subject or not, What matters is How I am leveraging my Interdisciplinary/Diverse Skills in Solving Customer Problems.
When Less of the Same is More
A person with Vague knowledge of different aspects of Chip Designing is much more productive than a bunch of Hyper Specialists who don’t care about what the hell is happening in the overall Project. Surviving with a Single Skill is a sure-shot way of building a disastrous career. No single Skill/Tool is sufficient in a complex, interconnected world where problems are usually multi-dimensional with multiple hidden layers. Generalists tend to become better at applying knowledge to a situation that they have never seen before, which is the essence of Innovation and Creativity.
How the Wicked World Was Made
VLSI is supposed to restrict factory-based Model of Working. Since Chip Design is more of Smart Work rather than Hard Work. In contrast, what we see today is there are too many departments in Semiconductor Companies than an SBI Bank or Paragon Factory. Each individual adding their unnecessary Complexity has created Complicated Patterns that are difficult to figure out. All the technical Jargon is becoming more difficult for us to correlate with underlying Basic Science & Math. Everyone is busy doing research that they don’t have time to know the way they are doing. Today’s Problems need multi-layered thinking which demands Versatile Skills and people who can dare to set their foot outside their World to search for solutions can only Survive in unprecedented situations.
Learning, Fast and Slow
More obstacles while Learning will make things more Frustrating, Slower, and Challenging which is not good in the short term but benefits us in the Long run, we call these “Desirable Difficulties”. We need to find a balance between when to learn Fast and when to learn Slow.
When we want our knowledge to be flexible, it has to be learned in extreme conditions and it is good to have mixed practice. Imagine an engineer who is trained under “Synthesis/STA/PD/DFT/PV Signoff” or “ASIC Design/ASIC Verification” they can easily differentiate a wide variety of Problems and can quickly arrive at Optimal solutions.
Thinking Outside Experience
Life is not so kind, there will be new problems that restrict you to probe solutions based on your Past Experience. This is the time where analogies will work for us. But to use analogies we need to improve our breadth of skills so that we can apply “known domain solutions” to “unknown domain problems” by recognizing similar patterns between multiple domains and be able to correlate unknown things with what we already know.
There are real-life situations where I took 3 days to solve an Unclocked registers problem, Where I messed up DFT clocks and Functional clocks. Had I known other Domain knowledge I could have saved Time/Money for Clients/Employers. We can improve our debugging/Quick-Problem-Solving skills by developing our breadth of Skills and then applying the solutions across domains. Your Willingness to dive into new domains will pay you unexpected Dividends in the Future!
The Outsider Advantage
I used to think it is Good to be an Insider than an Outsider. Was under the belief that people who are good in “local search” only can come up with Modern Innovation. Slowly I started realizing that Outsider Thinking can bring solutions that are more clever, cost-effective, and simple ( simple doesn’t mean easy!). I use the knowledge pocket gained in my old client location and reframe the problem in a better way to unlock better solutions for the current clients because I treat all the tangentially gained relevant experience as a single experience rather than Junk. So more Specialization/Uncertainty/Confusion increases in a Rapidly Changing World, it just gives an Unfair Advantage for a Generalist Since he/she can excavate Old knowledge and can weave it in a New Way.
Lateral Thinking With Withered Technology
Instead of getting lost in 28nm,14nm,7nm,5nm numbers it is good to play with Ideas and understanding the overall product, and gaining Domain knowledge. When you are consulting in Memory Design Company get enrolled in Memory Design courses, if you are in Networking Design company then listen to Computer Network Courses etc so that we don’t fall behind others. Never underestimate an Engineer who has a vague knowledge of everything! “Lateral and Vertical Thinkers” can excel even in cutting-edge technologies developed in VLSI Industry. Unfortunately, Many VLSI aspirants are happy to spend their career like a “Focused Frog” rather than a “Visionary Bird”. When are we going to get the Aerial View of the Problem? What are we doing to transition ourselves to a Visionary Bird?
Fooled by Experience
True Experience is working in environments where feedbacks are automatic and whenever an outcome comes out as a surprise how quickly we can adjust our ideas and come up with a radical approach. Personally, I would like to see engineers who work on RTL design/RTL Verification immediately swap their roles in the next Project, Same applies for backend Signoff/Implementation engineers to swap their roles at least once in a while to get the “Big Picture” and foster Creativity and Innovation in the Work we Do. So by this when we face Uncertain environments and wicked Problems, Breadth of experience can be invaluable.
When the patterns are similar and repetitive, narrow specialization can be remarkably productive. So don’t get fooled by your Expertise and Experience if so the results can be disastrous.
Learning to Drop Your Familiar Tools
Unless We start to drop our familiar Tools, things won’t get improved in the near future. Some of the statements which I hear from people around me are “This is not how we do here”, “ These are legacy paths”, “Let’s follow how it was done in Past projects let’s not take any risk” etc. Once we drop our familiar tools we are in process of unlearning and re-learning for flexibility and adapting to tough situations for the future. Dropping one’s familiar tools creates an “existential crisis” which motivates us to learn more to stay relevant in the Industry. Also, engineers should not feel if there is no data to back you up, Boss’s Decision is better than mine. We need to “use reason” when data is not sufficient to come to conclusion rather than removing the data points which don’t support our age-old Beliefs. Researchers found that Most effective Leaders had RANGE, they were, in effect, paradoxical, They could be demanding and nurturing, orderly and Entrepreneurial, hierarchical and individualistic all at the same time. A healthy ecosystem needs biodiversity. We need to be careful not to be too careful, this can limit our exploration mindset.
We need to preserve Range and inefficiency for a better VLSI ecosystem. Sometimes it is good to be amateurs and sometimes good to be experts all at once which can help us to dance across multiple disciplines. Cross-pollination of ideas should be respected.
Hyperspecialization is not good for Skill Development at least for developing countries like India where everyone brainwashes us to become Specialists.
I look forward to hearing from you Guys your opinions in the comments section. I wish everyone who reads this article will try to follow a few of the advice which are shared. See you at the Top. Jai Hind!
Linkedin Article Link:
