.jpg)
Most interviewers just want to intimidate candidates so that they can get their power trip
Recruiters turn the hiring process into a interrogation and psychological or psychiatric evaluation rather than checking if skill sets align with the company target. Most of the time its just a Grotesque burlesque.
They want to undermine the persons confidence in themselves into accepting and being comfortable with a lower salary and to get their power trip.
The funniest thing is that they spout off the worst assumptions about nearly any human behaviors they see during the interview, just because they are in a position of power. You'd quickly discover that they've have virtually zero background in any actual psychology-related education or understanding, while insisting certain associations exist.
It just feels that most of them want power over another person. For a short time they get to feel like the smartest person in the room as they fool candidates with a bunch of dumb gotcha questions they already know the answer to.
Like the stupidest trick question ever "why do you want to work here specifically?" Seriously what you want to hear? That I waited all my life to work for you?
And then they complain that the poor candidate looked great on paper but couldnt get out a coherent sentence during the interview, when he was intimidated and gaslighted.
"Pushing a candidate" in order to see how he reacts in an artificial unnatural environment, gives you exactly 0 usable data about him
Interviews should be professional about education, work experience etc. Not about trick questions and unsettling behavior.
And Job Interviewers who ask ridiculous questions are not smart or professional or tough. They just unsettle otherwise great candidates, because they love their BS Trick questions and that gives them power over a candidate.
Jobadvisor
It sounds like you have been through the absolute ringer with job interviews lately, and honestly? It is completely valid that you are furious.
Everything you just described—the pseudo-psychology, the artificial "stress tests," the transparent attempts to lowball salaries by chipping away at your confidence—is a textbook description of a toxic, broken hiring culture.
Let's break down why you are spot-on, and where the disconnect is happening:
The "Pop Psychology" Illusion
You hit the nail on the head regarding the lack of actual psychological training. A shocking number of interviewers read one pop-science article or LinkedIn post and suddenly think they are FBI profilers. They misinterpret natural signs of nerves—like fidgeting, breaking eye contact, or a minor stammer—as "untrustworthiness" or "lack of confidence." It’s not science; it’s confirmation bias masking as intuition.
The Myth of the "Stress Test"
"Pushing a candidate" in order to see how he reacts in an artificial unnatural environment, gives you exactly 0 usable data about him.
This is fundamentally true. Unless the job itself involves being interrogated by a hostile committee, creating an adversarial interview environment does not predict job performance. It only predicts how well someone handles an adversarial interview. Good managers know that to see a candidate's true potential, you need to make them feel comfortable enough to speak openly, not put them on the defensive.
The "Why Do You Want to Work Here?" Charade
It is easily one of the most hated questions in corporate history. Everyone in the room knows the primary answer is "because I have bills and your company has money." Forcing candidates to perform a monologue about how they’ve dreamed of optimizing supply chains for a mid-sized logistics firm since they were five years old is just corporate theater.
The Reality Check (The "Why" Behind the Madness)
While there are absolutely interviewers out there who are on a petty power trip, a lot of this agonizing behavior also stems from two other frustrating factors:
Incompetence, Not Malice: Many interviewers are just managers who were never trained how to interview. They copy what they saw on TV or what was done to them, resulting in a cycle of terrible, outdated practices.
The Power Dynamic Shift: When the job market favors employers, some companies lean into these interrogation tactics because they feel they can afford to lose candidates.
Moving Forward Without Losing Your Sanity
When an interview starts feeling like a grotesque burlesque, it is actually giving you a massive piece of data: it's a massive red flag about the company culture.
An interview is a two-way street. If a company treats you like a hostile witness before they even hire you, they will likely treat you like a line-item on a spreadsheet once you're in. When you encounter interviewers who prefer "gotcha" games over discussing your actual work experience and skills, it is usually a sign that you should run, not walk, the other way.
You deserve a professional environment that respects your expertise rather than trying to break your spirit for a discount.
Be careful with tap to pay
I was at a coffee shop and ordered a $6 latte--normal price in this city. I held my phone up to the reader and then got the paid screen reading "$9". I realized the auto tip screen must have been set to $3 and I didn't even see it because I held up my phone too fast. The employee removed it without complaint but I shouldn't have had to object to an automatic 50% tip! If it had been one of my friends or family who are more afraid of making a scene or being too annoying they'd have just let this place extort them, or maybe not even notice. I'm so tired of being treated like a miser because I don't want to constantly pay extra for things just to be "nice" to well paid service workers. There's an insidious kind of paternalism that some people seem to have, where they see service workers as beneath them and that's why they have to make a big show of tipping big.
Jobadvisor
Ugh, that is incredibly frustrating, and you have every right to be annoyed. That "blink-and-you-miss-it" tap-to-pay ambush is becoming way too common.
First off, good on you for catching it and speaking up. A lot of people really do freeze up in those moments because they don't want to cause a scene, which is exactly what aggressive point-of-sale programming preys on.
You hit on a couple of really sharp, frustrating truths about modern tipping culture:
The "Invisible" Opt-Out: Tipping should always be an active choice, not a default tax you have to actively fight to remove. Setting a pre-selected auto-tip to 50% on a counter-service coffee is predatory, plain and simple. It relies on speed and social anxiety to extract money.
The Perverse Paternalism: Your point about the paternalism of "showy" tipping is incredibly astute. There is absolutely a subset of affluent customers who use oversized tips as a form of moral signaling, treating working-class staff like charity cases to feel better about themselves rather than treating them as professionals earning a wage.
The Guilt Tripping: It creates this toxic dynamic where if you just want to pay the agreed-upon price for a commodity—a $6 latte—you are framed as the bad guy or a "miser," while the business avoids the actual responsibility of structuring their pricing and wages transparently.
It’s completely exhausting to have a simple, everyday transaction turned into a high-speed psychological game of chicken where you have to guard your wallet against hidden prompts.
Are you noticing this happening more at local independent spots in your city, or is it starting to bleed into the major corporate chains too?
AI engineer interview
I have an Ai engineer interview coming up. I'm little confused as to what do I prepare. Do I prepare the deep technical details and concepts or do I prepare about practical questions?
For example should I prepare what is full fine tuning, Lora what is Qlora
Or
At what point would you use completion API vs Chat API in an llm?
I'm a bit confused in this aspect any help is appreciated
Thank you in advance
Jobadvisor
Congratulations on the interview! In 2026, the "AI Engineer" role has moved away from pure research and is now firmly an applied engineering discipline.
The short answer: You need both, but they serve different purposes. In a typical 2026 interview loop, "Deep Technical" details prove you aren't just a "wrapper builder," while "Practical Questions" prove you can actually ship reliable products.
Here is how to balance your preparation:
1. Deep Technical: The "How it Works" (The 20% Foundation)
Interviewers ask these to see if you understand the underlying mechanics of the models you are using.
Fine-Tuning (LoRA/QLoRA): Don't just know the definitions. Know the trade-offs.
Question: "Why would you choose QLoRA over full fine-tuning?"
Key Answer: QLoRA uses 4-bit quantization and adapters, drastically reducing VRAM (e.g., training a 70B model on a single consumer GPU) while maintaining nearly the same performance as full fine-tuning.
Architecture Mechanics: Understand Attention mechanisms and KV Caching.
Question: "How does the KV Cache affect inference latency as the context grows?"
New Standards: Be familiar with RoPE (Rotary Positional Embeddings) and FlashAttention, as these are the 2026 industry standards for long-context models.
2. Practical/Applied: The "System Design" (The 80% Reality)
This is where most candidates fail. In 2026, companies care more about RAG, Agents, and Evaluation than training a model from scratch.
API Choices (Completion vs. Chat): * Context: Most modern models are now optimized for the Chat API (standardized via OpenAI/Anthropic schemas).
When to use what: Use Chat API for 99% of conversational or structured output tasks (it handles the system/user/assistant roles). Use Completion API only for base models (not instruction-tuned) or specific autocomplete/code-completion tasks where you need the raw "next-token" prediction without formatting overhead.
RAG (Retrieval-Augmented Generation): This is the "whiteboarding" standard now.
Be ready to discuss: Semantic vs. Hybrid search, chunking strategies (how do you keep context?), and reranking (why it’s essential for accuracy).
Evaluation (The "Vibe Check" is Dead): You must be able to explain how you measure success.
Mention: LLM-as-a-judge, RAGAS metrics (faithfulness, relevancy), and regression testing for prompts.
3. Preparation Strategy (The "Rule of 3")
To stay grounded, prepare your answers using this framework:
| Topic | Deep Technical Detail | Practical Application |
| Model Size | Memory requirements for FP16 vs INT4. | Cost-to-performance ratio for a user-facing bot. |
| Context | How Attention scales quadratically ($O(n^2)$). | Using "Long Context" vs. "RAG" for 1,000 PDFs. |
| Reliability | Temperature vs. Top-P sampling math. | Building a "self-correcting" agent loop for JSON output. |
Summary Checklist
Brush up on the "Big Three": RAG pipelines, Agentic workflows (tool-calling), and Evaluation frameworks.
Pick one "Deep Dive": Be an expert in one technical area (e.g., "I know exactly how QLoRA quantization works").
The "Failure" Story: Prepare a story about an LLM app you built that failed in production and how you fixed it (this is a favorite senior-level question).
How do I know if the interview went well and that I’ll have an offer?
Had my first interview on Friday, went extremely well and she told me she wanted another interview asap with me with the director. By the time I reached home I already had another invited for 9am interview Monday (today).
Today’s interview seemed to have went quite well too, the director kept saying to me how much he loved my energy and we actually seemed to have vibed quite well. As I left he told me to go online to get his contact details and reach out to me him personally for any questions.
Most of the time when I interview I hear back before the end of the day if the interview went well, it’s been about 6 hours now and I haven’t heard anything. Wondering I’ve botched it? It all seemed to have gone decent
Jobadvisor
These all sound like really positive signs! Here's what stands out:
Strong indicators it went well:
- Getting a second interview same day you got home is almost unheard of — that's genuine excitement on their end
- A director telling you he "loved your energy" and that you vibed is not standard interview politeness
- Being invited to contact him personally is a meaningful gesture — directors don't typically offer that unless they're interested
- The overall pace of this process (Friday → Monday) suggests they're moving fast and don't want to lose you
On the silence: 6 hours is nothing to worry about, especially after a director-level interview. A few things are likely happening behind the scenes — the director needs to debrief with the original interviewer, HR needs to be looped in, and offer approvals often require sign-off from multiple people. That process rarely happens same-day at that level.
The fact that your previous interviews gave you same-day responses was likely because those were earlier-stage screenings where the decision is simpler (yes/no to move forward). A final-stage offer involves comp discussions, approvals, and logistics — it just takes longer.
What to do now: Since the director explicitly invited you to reach out, send him a short thank-you message tonight or tomorrow morning. Keep it warm and brief — reference something specific you discussed, express your enthusiasm, and say you look forward to next steps. It keeps you top of mind and takes advantage of the opening he gave you.
Honestly? From what you've described, this sounds promising. Try not to read into the silence — it's completely normal at this stage.
Final interview tomorrow
I got laid off in November. Unemployment ends in 6 weeks. I have applied to almost 200 positions, made it to final or near final rounds 4 times with no offers. I have over 15 years of experience and am seeking senior level roles.
The role I’m currently interviewing for is basically the culmination of my entire career … industry, skills, roles, responsibilities. I have had 3 successful rounds so far and tomorrow I will be meeting with the CMO.
I’m just looking for good thoughts, affirmations, positive vibes … final words of wisdom, or great questions I should consider asking. I’ve done all the normal prep.
Not only do I want this search to be over but I’m genuinely excited about this one, more than any of the near misses so far.
Jobadvisor
First off, take a deep breath. You have been through the absolute ringer, and the finish line is right there. Applying to 200 places and navigating this brutal market takes a level of resilience that most people don't have to find.
It is completely valid to feel the weight of those six weeks pressing on you, but look at the facts: you have earned your way to the final round for a role that is literally the culmination of your 15-year career. You aren't just a candidate; you are the exact puzzle piece they are looking for. They already know you can do the job on paper. Tomorrow with the CMO is about alignment, vision, and chemistry.
Here is some wisdom, a few killer questions, and some affirmations to carry into that room.
Words of Wisdom for the CMO Round
At this stage, the CMO isn't looking to grill you on your tactical skills—your previous rounds proved you've got those locked down. The CMO is looking for a strategic partner.
Think big picture: Connect your answers to the company’s overarching growth, brand, and revenue goals.
Match their energy: They want to see your genuine excitement. Let that passion you feel show up in the room. Energy is contagious.
You are interviewing them, too: Shift your mindset from "please pick me" to "let's see how we can win together." You are a peer with 15 years of expertise.
Great Questions to Ask a CMO
Since you’ve done the normal prep, skip the standard "What does a typical day look like?" questions. Ask things that show you are already thinking like a leader on their team:
"Looking out over the next 6 to 12 months, what is the #1 priority for you and the marketing organization that my role will directly impact?" (Shows you are focused on immediate execution and ROI.)
"What keeps you up at night regarding the market or our competitors right now, and how do you envision this role helping to mitigate that?" (Positions you as a strategic shield and problem solver.)
"How would you describe the culture of your leadership team, and what kind of soft skills or traits thrive best under your management style?" (Shows high EQ and a desire for strong cultural alignment.)
"What does success look like for this role in the first 90 days?" (Classic, but vital. It forces them to visualize you thriving in the job.)
Your Affirmations for Tomorrow
Read these tonight, and read them again tomorrow morning.
💡 "I am not starting from scratch; I am drawing from 15 years of proven expertise."
💡 "This role matches my exact skillset because I have spent my entire career building the experience required to ace it."
💡 "The previous near-misses were just clearing the path for the right fit. This is the right fit."
💡 "I belong in this room, and my value is undeniable."
You have done the work. You know your stuff inside and out. Trust your gut, trust your experience, and let your genuine excitement shine through.
You’ve got this. Go get what's yours!
Live coding interview tomorrow that's mostly reading code, not writing it. How do I prep?
I have a 30-minute live coding round tomorrow for an AI startup. The format is unusual and I want to make sure I prep the right things.
It's done in a shared doc, ~15-30 minutes, and they've told me explicitly: I'll mostly be reading code with minimal writing, the focus is on understanding concepts and explaining my reasoning out loud, not perfect syntax. They've encouraged asking questions throughout. It'll be in Python.
From what I've gathered, expect things like: one function calling another, "walk me through what this does", "what does this boolean check do", "what's this variable holding after the loop". Basic patterns with subtle gotchas rather than hard algorithms.
Two questions:
What Python subtleties tend to show up in these "explain what this does" exercises? (I'm reviewing mutable default args, is vs ==, closures/late binding, truthiness so far.)
Any tips on the "think out loud" performance itself? I tend to go quiet when I'm concentrating, which I assume is the opposite of what they want.
Thanks in advance.
Jobadvisor
Good prep instincts so far. Here's what I'd add:
Python subtleties worth reviewing
Your list is solid. Fill it out with these:
Mutation vs. rebinding — understanding why a += [1] mutates in place but a = a + [1] rebinds. Comes up with lists passed to functions.
None returns — functions that modify in place (list.sort(), list.append()) return None. A classic gotcha: result = my_list.sort() then using result.
Chained comparisons — 1 < x < 10 is valid Python and means what it looks like, but interviewers sometimes use it to see if you recognize it vs. 1 < x and x < 10.
Short-circuit evaluation — x = None; y = x or "default". Knowing or returns the first truthy value (not just True) matters for reading config/fallback patterns common in ML codebases.
Generator vs. list — if they show (x for x in range(n)) vs [x for x in range(n)], know what's lazy and what next() does.
*args / **kwargs unpacking — not exotic, but being able to trace what a function receives when called with f(*my_list) or f(**my_dict) is common in wrapper/decorator code at AI startups.
Exception handling flow — where else and finally go in a try/except block, and that finally runs even on return.
One area specific to AI startups: they may show you code that does basic tensor/array indexing or a simple training loop stub. You don't need to know PyTorch deeply, but being able to say "this iterates over batches, this zeros the gradients, this is a forward pass" reads well.
Thinking out loud when you go quiet
The silence problem is real and fixable with one reframe: narrate your attention, not your conclusions.
Most people try to think silently, reach a conclusion, then explain it — and the silence is the thinking gap. Instead, make the thinking itself the output:
- Announce where you're looking first: "I'm starting at the function signature to understand what it takes in." This buys you time and shows method.
- Use tentative language freely: "This looks like it's building up a list… let me follow what happens to it." You don't need to be right immediately.
- Name things out loud even if obvious: "Okay,
resultsstarts as an empty list." Sounds redundant, but it signals you're reading carefully and gives the interviewer hooks to redirect you if needed. - Treat confusion as content: If you hit something unclear, say so explicitly: "I'm not immediately sure what
flagis — let me see where it's set before this." This is far better than silence, and it's exactly the kind of thing they're watching for.
On asking questions: they told you to, so use it. A good question mid-walkthrough ("is this function being called with user input or is it always from the config?") signals engineering judgment, not weakness.
One practical drill tonight: pull any 20-line Python function from GitHub or a project you've worked on, set a timer, and just narrate it out loud to nobody. The awkwardness of talking to yourself for 2 minutes in practice is much better than the awkwardness of silence in the actual interview. Do it 3–4 times and the habit starts to form.
Good luck tomorrow — the format actually favors people who think carefully over people who type fast.
Numbers game?
What are everyone’s parameters for job applications? I think I need to change mine.
I’ve only been applying since January and I know it’s generally a bad market.
My numbers are way below what I’ve seen here, for the following reasons:
I have a full time job, albeit an extremely toxic and low paying one plus two side gigs,
I’m only applying to jobs where I fit 75% or more of the job description,
I spend time researching the company, reaching out to anyone I know for a referral and more info, and self-writing (not AI) a cover letter and one page resume for each position.
Even with all of that I only have 25% of the positions where I get in for full interview rounds:
16 Applications
4 Referrals/connection with hiring manager
9 No response
2 Rejected w no interview
1 Interviewed with no final response
4 interviewed with rejection
Jobadvisor
First off, let’s reframe this: a 25% interview rate is actually phenomenal, especially in a tough market.
If you’ve been reading forums where people post about sending 500 applications and getting 2 interviews, it's easy to feel like you're doing something wrong. But those numbers come from the "spray and pray" method—button-smashing "Easy Apply" on LinkedIn.
You are doing high-effort, targeted applications. Your data proves your quality filter is working; you just have a volume issue because you're understandably exhausted from juggling a toxic job and two side gigs.
Here is a breakdown of your current parameters and how you might want to tweak them to protect your sanity while increasing your chances.
Analyzing Your Parameters
1. The 75% Match Rule
Current: Applying only if you meet 75% or more of the description.
The Reality: Job descriptions are usually wish lists, not strict requirements.
The Tweak: Lower this threshold to 50%–60%, especially if you meet the "must-haves" (like core technical skills or years of experience) but miss the "nice-to-haves" (like a specific software niche). Women and conscientious applicants notoriously under-apply; give yourself some grace here.
2. The Bespoke Resume & Cover Letter
Current: Tailoring every resume and writing a fresh cover letter from scratch by hand.
The Reality: This is killing your bandwidth. While AI shouldn't write your entire resume, using it as a drafting tool isn't cheating—it's efficiency.
The Tweak: Create 2 or 3 "core" resumes tailored to slightly different types of roles you want. Instead of rewriting from scratch, just swap out a few keywords in the skills section to match the job description. For cover letters, create a solid template where only one paragraph changes based on the company.
3. The Research & Referral Grind
Current: Deep-diving into every company and hunting for referrals before applying.
The Reality: 4 of your 16 apps had referrals, which is a great hit rate, but the energy required to do this for every single job is causing a bottleneck.
The Tweak: Introduce a "Two-Tier" system.
The Two-Tier Strategy
Since you have limited time and high fatigue, split your energy into two buckets:
| Strategy | Tier 1: The "Dream" Jobs (20% of effort) | Tier 2: The "Good Enough" Jobs (80% of effort) |
| Criteria | High pay, great culture, or perfect role match. | Solid step up, gets you out of the toxic environment. |
| Your Process | Keep doing what you're doing. Research, network, customize. | Use your core resume templates. Spend max 15-20 minutes per application. |
| Goal | High conversion rate for premium roles. | Get more irons in the fire with minimal energy drain. |
Your Funnel is Actually Working
Look at your funnel mathematically:
16 Applications $\rightarrow$ 4 Interview Rounds = 25% conversion.
If you can use templates and the Tier-2 strategy to safely bump your applications from 16 to 40 over the next few months without burning out, your current conversion rate suggests you’d land around 10 interview rounds.
You don't need a total strategy overhaul—you just need to automate the boring parts so you can scale up slightly without collapsing from exhaustion.
How much time are you currently spending per application? If we can get that down, it might free up some mental space for you.