How to prepare for career in embedded + AI intersection ?

Hi. I have around 8 years of experience working in India, and most of that is around process analysis for mfg processes, doe analysis, causality etc. and I have also worked on RAG, data engineering for LLMs.

Off-late I am interested in the intersection of embedded systems and AI (SLMs, localised training and inference at edge, robotics etc.). Part of the interest also stems from the fear that with most of my day job slowly transitioning to talking in broken english to Claude, I feel this is a niche domain where there is still lot of room to create and build.

Need some guidance from experts here how to build my portfolio, what courses to take, how to pick some pet projects to work on to build my GitHub profile etc. How can I build expertise in this hardware+ai domain ?

Thanks in advance.

Jobadvisor

Great instinct — the embedded+AI intersection is genuinely underserved right now, and your background is actually a stronger foundation than you might think. Process analysis, DoE, and causality work maps directly onto model evaluation and optimization on constrained hardware. Here's a structured path:

Ground yourself in the hardware layer first

Before touching AI on edge, you need fluency with the substrate. Start with:

• Microcontrollers: Get a Raspberry Pi 5 and a STM32 or ESP32 dev board. The Pi for Linux-based inference, the MCU for truly bare-metal work.
• C/C++ refresher: Most edge AI runtime code (TFLite Micro, ONNX Runtime, llama.cpp) is C/C++. You don't need to be a systems programmer, but you need to read and modify this code comfortably.
• RTOS basics: FreeRTOS is the lingua franca. Even a surface-level understanding of task scheduling and memory management matters when you're squeezing a model onto 512KB RAM.

The core technical stack to learn

Quantization and model compression — This is the central skill. INT8, INT4, GPTQ, AWQ. Understand what happens to accuracy/latency tradeoffs. Hugging Face + bitsandbytes is the easiest entry point; then go deeper with TensorRT or ONNX.

Edge inference runtimes — In rough order of importance:

• llama.cpp (LLMs on CPU/edge, very active community)
• TensorFlow Lite Micro (for MCU-class devices)
• ONNX Runtime with execution providers
• ExecuTorch (Meta's PyTorch-native edge runtime, rapidly growing)

Hardware-aware profiling — Your DoE background is a superpower here. Systematically benchmarking latency, memory, power consumption across quantization levels, batch sizes, and hardware configs is exactly DoE work. Most ML people do this sloppily; you won't.

Portfolio projects that will actually stand out

The goal is projects that show both sides of the stack — not just "I ran a model on a Pi."

1. Quantization benchmark suite — Take 3-4 SLMs (Phi-3 Mini, Gemma 2B, Qwen2-0.5B), run them across INT8/INT4/FP16, measure latency + perplexity + RAM on a Pi 5. Write it up like a proper DoE study. This is immediately useful to hiring managers and very rare on GitHub.

2. Sensor-to-inference pipeline — Connect a sensor (vibration, thermal, audio) to a Pi or ESP32, run a small anomaly detection or classification model locally, and log results. Manufacturing background makes this authentic — frame it as predictive maintenance.

3. RAG on the edge — Given your RAG background, build a local RAG system using llama.cpp + a vector store (ChromaDB or sqlite-vec) that runs entirely offline on a Pi. Document the memory constraints you hit and how you solved them.

4. Robotics entry point — ROS 2 + a small differential drive robot + a vision model (MobileNet or similar) for object detection. Robotics is exploding and very few people have both the software and analytical chops.

Courses worth your time

• Fast.ai Practical Deep Learning — still the best for intuition, not math-heavy
• TinyML and Efficient Deep Learning (MIT 6.5940, free on YouTube) — directly on point
• CS144 or similar for networking basics if you want to work on fleet/OTA update systems for edge devices
• Coursera: Fundamentals of Embedded Systems (U Colorado) if you need the structured MCU foundation

Skip most "AI for IoT" courses — they're usually shallow. Go to primary sources (papers, runtime documentation, GitHub issues).

How to position your existing experience

Don't hide the manufacturing/process background — reframe it:

• DoE experience → systematic hardware benchmarking
• Causality/process analysis → root cause analysis for model degradation on edge
• RAG/data engineering → on-device retrieval, knowledge distillation pipelines
• 8 years in India's manufacturing sector → industrial IoT and predictive maintenance use cases

This combination is genuinely rare. Most edge AI people come from pure EE or pure ML; very few have your applied industrial + LLM background.

Realistic 6-month plan

Month	Focus
1-2	C++ basics, FreeRTOS, get llama.cpp running on Pi, first benchmark project
3-4	TFLite Micro on ESP32, quantization deep dive, sensor pipeline project
5-6	RAG-on-edge project, start ROS 2, write up everything as proper technical blog posts

The blog posts matter as much as the GitHub repos — write them like engineering reports, not tutorials. That's what differentiates someone with your background.

28 years old. Good savings. No idea how to progress my life. What should I do?

Not sure if this is the right place to post this.

Basically I’m a 28 year old man, engaged to a 27 year old woman. We have collectively saved over 100k USD. We live in a low tax state in the south in a major city. We work as a waiter and nanny, refurbish furniture on the side - our savings come from a kind of fluke job position we got together, and then a very lucky investment that profited me 50k.

I have basically no work experience. Lived with my parents until I was 23 (battled a severe digestive illness for my early 20’s, led to me not pursuing college) and I’ve been a waiter, private chef, homeschool teacher, delivery driver… nothing of consequence.

Living with my parents was helpful financially but completely mentally destructive - they’re all mentally ill and got foreclosed upon during my stay. We both have zero financial help from any family and likely never will. Almost no family to lean on.

We both only have high school diplomas, I might have a semester or two of community college from ten years ago- we’re quite intelligent but have had rough upbringings - lots of trauma and crazy scenarios that took time away from being able to focus on ourselves.

We both struggle with mental health issues, very bad ADHD and I also have struggled with severe depression since I was a child. We’re both sober. We’re very frugal. We’re trying to figure out what on earth to do so that we can get a step up and improve our quality of life. I feel somewhat humiliated at my inability to progress myself. The things I’m good at (music, writing) are incredibly difficult to monetize (or appear to be on the surface). I want to have children in a few years and not have my wife working as much as I am.

If you were in your late 20’s with 100k in savings in the US, what would you do?

I own a vehicle, an iPhone, a laptop - live in America. Am relatively young. It would seem that there’s a potential for me to do many things. But my brain just can’t push itself into picking something to pursue for long enough. Im constantly bouncing.

My self worth is, naturally, very low. I am charismatic and somewhat conventionally attractive but my low self esteem and terrible focus has led me to making virtually no business connections or friendships of financial/ career consequence. I think I’m a fantastic musician (don’t we all) yet have no one to help me do it. The thought of doing it alone makes me pull back from it, as I have very little time, it’s a vicious cycle.

Interests: music, writing music, recording music, poetry, screenwriting, film in general, travel, the outdoors, animals, nature, natural health, religion and spiritualism, philosophy, history, etc.

Not opposed to starting a small business.

Jobadvisor

First off, take a deep breath. You are $100\text{k}$ ahead of the vast majority of people your age, you are sober, you are engaged to a partner you build a life with, and you survived a chaotic upbringing and a severe chronic illness.

That isn't "nothing of consequence." That is a massive triumph.

The humiliation you feel is just your low self-esteem playing tricks on you because you're comparing your inside chaos to everyone else's polished outside. Let’s look at your situation objectively and lay out a practical, ADHD-friendly roadmap to get you where you want to be.

1. Protect the Fortress (The Financial Foundation)

Do not touch that $100\text{k}$ to fund a "fringe" dream or a high-risk business right now. That money is your safety net, your peace of mind, and the future down payment on a house for the family you want to build.

• Park It: If you haven't already, put the bulk of this money into a High-Yield Savings Account (HYSA) or a ladder of CDs. At current rates, it should be passively making you hundreds of dollars a month completely risk-free.

• Keep Living Lean: Your frugality is your superpower. Maintain your current lifestyle while you transition so you don't burn through your capital.

2. Pivot Careers (Without 4 Years of College)

You need a career that accommodates ADHD (varied tasks, active, or project-based) and offers a higher income ceiling than waiting tables. Given your interests in nature, the outdoors, and your hands-on experience refurbishing furniture, here are a few directions:

• The Trades / High-End Craftsmanship: You already refurbish furniture. Have you considered looking into high-end carpentry, cabinet making, or upholstery? These industries are desperate for young, intelligent people, and they pay incredibly well once you learn the craft.

• Green Industries / Outdoors: Look into park management, surveying, or certified arborist work. They get you outside, involve natural science, and often require certifications rather than a full 4-year degree.

• Digital Content / AV Tech: Because you love music, recording, and film, look into the boring but lucrative side of the industry: Corporate Audio/Visual (AV) tech, live sound for events, or video editing. You can get certificates or entry-level gigs that pay much better than waiting tables, keeping you close to your passions without relying on them for rent.

3. Monetize Your Creativity Collaboratively

You mentioned you are a fantastic musician but pull back because doing it alone is daunting. ADHD brains thrive on body-doubling and external accountability.

• Don't try to make it as a solo rockstar. Look for local indie filmmakers, podcasters, or theater groups who need scoring, sound design, or foley work.

• Use your charisma. Go to local open mics, film meetups, or community theater builds. Don't go to "network" for money; go to find your people. When you collaborate, the pressure leaves your shoulders.

4. Tackle the ADHD and Self-Worth

You can have a million dollars in the bank, but if your brain is constantly bouncing and telling you that you aren't good enough, you will feel stuck.

• Treat the ADHD: If you aren't medicated or in ADHD-specific coaching, make this your top priority. Investing a couple of thousand dollars of your savings into proper mental health care (therapy, executive function coaching, or psychiatric evaluation) will yield a 1000% return on investment.

• Reframe your past: You didn't "waste" your early 20s. You were surviving. You are now entering your prime with wisdom that your peers who had easy lives don't possess.

A Note for the Next 3 Years: Your goal is to get your wife out of full-time work and start a family. Focus the next 24 months entirely on stability and skill acquisition. Pick one direction for work—even if it's just a 2-year trade program or building your furniture business into a legitimate LLC—and commit to it for just 12 months to see it through.

You have a clean slate, a massive financial cushion, and a partner by your side. You are not starting from zero; you are starting from experience.

Which of those career paths (trades/craftsmanship, outdoor/nature industries, or audio/visual tech) feels like something you could see yourself doing day-to-day?

Leaving with no notice and I dont feel bad about it?

I worked at a small architecture firm for six years. Started as a junior and worked my way up to running most of the client facing projects on my own. I genuinely loved the work and the people I worked with and even though the pay wasnt amazing I stayed because the environment was good and I felt like I mattered there.

A year ago my mom got diagnosed with something serious and I had to start going to appointments with her during the week because my dad cant drive and my sister lives in another country. I talked to my boss about it ahead of time and asked if I could shift my hours around on appointment days and make up the time later in the week. He said sure no problem well make it work.

I started noticing little comments from the office manager like must be nice having a flexible schedule or I wish I could just leave in the middle of the day even though my work was getting done on time every single time.

Then my mom had a bad reaction to her treatment and I had to take three days off to be with her. I called my boss and he said take the time you need. I had PTO and I used it.

When I came back there was a formal meeting with HR waiting for me. They said my attendance was a pattern of concern and I needed to commit to full time in office or theyd reconsider my role. My boss was sitting right there and didnt say a word. The same guy who told me take the time just sat there staring at his hands.

I didnt argue. I nodded and went back to my desk. Started looking that same night and found something within a month. Better pay closer to home and they didnt even blink when I mentioned my moms situation in the interview.

Im using up my PTO now and friday is my last day. No notice.

Jobadvisor

What you're describing isn't impulsive — it's a completely proportionate response to a specific betrayal.

You didn't just leave because of a bad policy or a stressful workplace. You left because your boss looked you in the eye, told you to take the time you needed, and then sat silent in a meeting that used that time against you. That's not a misunderstanding. That's a person choosing the path of least resistance at your expense, after six years of you building their client relationships.

The "no notice" framing makes it sound reckless, but you actually gave them a year of good faith. You communicated proactively, you kept your work on time, you used PTO you earned. They responded by treating a cancer diagnosis as an attendance pattern. Notice is a professional courtesy — it's not owed unconditionally, and it's especially not owed to people who've already decided your personal crisis is an HR matter.

The office manager's comments are worth naming too, because that kind of low-grade resentment — must be nice having a flexible schedule — is what poisons workplaces quietly. Someone was building a case, and your boss either didn't notice or didn't care enough to stop it.

The fact that your new employer didn't blink when you mentioned your mom's situation isn't a small thing. That's what a baseline decent workplace looks like. You found out what this one actually was when the pressure came.

You don't owe guilt to people who chose paperwork over you while your mom was sick. Take care of her, start the new job, and let Friday be Friday.

Supervisor stole my idea what now?

Hello! Been at my company for 3+ years. Good at what I do. Got a new supervisor who knows nothing about what we do at my company. Just got the role bc they have supervisor experience. Basically I want that role but they hired them before I made it known I was interested in the role. I didn’t know this person would be joining my particular team so I never thot it was a big deal they existed. Anywho - I pitched an idea to this supervisor months ago - had a gut feeling they would steal so I put it in writing to the director - whom brushed it off. Fast forward to now, just got out of a meeting where the director was giving this supervisor kudos for his idea. It was my idea. 100% came from my mouth and he is taking credit for it. What’s even better is that since he doesn’t know anything about our company and how to do the job, I am being tasked with doing 100% of the training material, I then have to train them and then I get cut from the actual meetings where they will be presenting this material to the rest of my coworkers.

TLDR - pitched an idea, it got stolen, now I have to train the person who stole it on how to do the project bc they don’t know anything about what they do or how to do this project because it was not their idea.

I am contemplating responding to the original email thread where I pitched this idea and say “I’m excited to get working on this project. I had not heard anything about this progressing until today and I am excited to see what (insert supervisors name) does with it.”

How would yall go about this? I want to know how am I going to get credit for this project at year end review when it was stolen from me.

Jobadvisor

That email reply idea is actually pretty good instincts — passive enough to not blow things up, but it puts your name on record in the thread. Let me help you sharpen it and think through the broader situation.

On the email

Your draft is close but a little too cheerful — it reads as slightly sycophantic toward the supervisor. Try something more neutral that just establishes your ownership clearly:

"Happy to see this moving forward — this is the project I pitched back in [month]. I'll get started on the training materials. Should I send updates directly to you or copy [supervisor]?"

That version does three things: timestamps your idea, signals you're the one doing the work, and subtly highlights that you're the one who actually knows how to execute it.

On the year-end review problem

Start building a paper trail now, not later. Every deliverable you create for this project — training docs, materials, outlines — send it via email rather than handing it over verbally. "Here's the first draft of the training module" in writing is evidence. A conversation in a hallway is nothing.

Also document the original pitch email somewhere personal (screenshot, forward to a personal address) in case anything gets lost or threads get buried.

The harder truth

The director already brushed you off once, so going back with "he stole my idea" without new evidence is likely to go nowhere and could make you look difficult. Your leverage here isn't the idea anymore — it's the execution. You're the one building this. Make that undeniable and visible.

If you want the supervisor role eventually, the play is demonstrating that you're the operational backbone of this project while he's the face of it — and making sure people one level above the director can see that clearly over time.

What's your relationship like with the director directly?

What career offers the highest chance of becoming wealthy at a young age?

I’m currently a high school student planning to pursue a STEM career. I come from a low-income family, and I always been longing for earning a lot of money after graduating college.

With that in mind, which career (not necessarily STEM. Just ANY career) offer the highest potential for building significant wealth? By “significant wealth,” I mean becoming a millionaire and ideally reaching a net worth of around $5–6 million as early as possible, perhaps by my early 30s if that’s realistic.

Jobadvisor

Honest answer: the paths that get people to $5-6M by their 30s are real, but they're narrow and the selection effects are brutal. Here's the actual landscape:

The highest-ceiling paths

Software engineering → equity at a startup This is the most reliable repeatable path. A senior SWE at a top tech company (Google, Meta, etc.) earns $300-500K+ in total compensation within 5-8 years of graduating. That alone gets you to $1-2M relatively safely. The lottery ticket on top is joining an early-stage startup for equity — if it exits, $5M+ is realistic. Most startups fail, but the base salary floor is high enough that failed bets don't ruin you.

Founding a company The highest EV path but also the highest variance. Most founders fail multiple times. The ones who hit big usually had domain expertise, timing, and some luck. Your STEM background is a good foundation — technical founders have an advantage over pure business founders.

Finance — specifically quant trading or PE/VC Quant traders at top firms (Jane Street, Citadel, Two Sigma) earn $500K-$2M+ within a few years. The catch: these firms hire from a very small pool of top math/CS graduates from elite universities, and the interview process is extremely competitive. Private equity has a similar profile — high ceiling, very narrow door.

High-income STEM → disciplined investing Less glamorous but more reliable than people admit. A doctor, dentist, or pharmacist earning $250-400K who lives modestly and invests aggressively can reach $5M by mid-30s to early 40s. Slower, but the probability of getting there eventually is much higher than startup equity.

What actually separates people who get there

• Equity over salary — Salaries make you comfortable; ownership makes you wealthy. Whether that's stock options, a business, or real estate, wealth compounds through ownership.
• Location and company tier matter enormously — A SWE at Google earns 3-4x what the same person earns at a regional company. The delta is massive early in your career.
• Lifestyle inflation is the killer — Most people who earn $300K spend $280K. The ones who build wealth keep living like they earn $80K for a long time.
• Network — Especially for startups and finance, who you know determines what opportunities you see.

Realistic calibration for your situation

Coming from a low-income family is genuinely a disadvantage for the finance/VC path (network matters a lot there) but less so for software. The SWE → top tech company → startup equity path is the most meritocratic of the high-ceiling options and the most accessible regardless of background.

If you're strong at math and CS, aim for the best university you can get into (scholarships exist), study CS or math, intern aggressively, and target the top-tier tech companies first. That's the most reliable on-ramp.

$5-6M by early 30s is possible but not common even among high earners — being honest with you about that. $1-2M by 35 is much more achievable on the SWE path if you're disciplined. The jump to $5M usually involves either a successful exit or a decade of high earnings plus smart investing.

What subjects are you strongest in right now?