Jugal Bhatt (@jugaldb) header
Jugal Bhatt (@jugaldb) profile picture

AI & Tech Content Creator | Hackathon Judge | Speaker | Software Engineer @ Amazon | UIUC CS Grad 2025

San Francisco, CA
Joined September 2015
Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)Photo by Jugal Bhatt (@jugaldb)
Freshers do be collecting clubs like infinity stones
5
6
98
This George Russell hate watch from my entire timeline might be one of the best things that has happened in F1
1
119
Hi X, I'm Jugal. One of my first posts here, so here's the journey: 2019: Started CS at @VIT_univ 2022: Joined @FamAppByTrio 2024: Started my master's at @UofIllinois, for free 2025: Graduated. Founding engineer at @LiteLLM, moved to SF, built their MCP Gateway from scratch 2025: Joined @amazon Now: AI engineer at Amazon I write Ascend, a weekly Substack on AI and tech. If you're breaking into tech or AI, let's connect.
26
1
64
7,102
If I had invested all the money on @AnthropicAI IPO I have spent on subscriptions instead of using it for Claude code. . . . . . . I’d be poorer than I’m now
1
120
Hot take: 90% of "AI engineers" on LinkedIn are just prompt wrappers with a new job title. Calling an LLM API is not AI engineering. AI engineering is: Building evals before features. Hunting the failure cases. Catching the model when it's confidently wrong. Anyone can ship a demo in a weekend. Almost nobody can ship one you'd trust with real users.
2
3
200
Stop grinding 500 LeetCode problems. I did 120 and cleared Google, Meta, and Amazon. Problem count was never the bottleneck. Pattern recognition is. If you can't name the pattern from the problem statement alone, problem #501 won't save you. Grinding feels productive. That's the trap. I wrote down exactly how I did it: jugaldb.substack.com/p/i-jus…
2
145
Lenny's new job market report has a stat I can't stop thinking about: US engineering openings: +44% US engineering hires: −17% Both since Jan 2024. Inspired by @lennysan's newsletter, I decided to dig deeper: 10-Ks, BLS data and layoff trackers, to find where the jobs went 🧵
1
114
10/ The playbook by role: Engineers → Move toward the model or the customer. AI/ML, infra and customer-facing roles are gaining share → Front-end or mobile only? Reposition as full-stack or product engineering → Grow your scope. Staff pay rose 7.5% in 2025 vs 1.6% at entry level PMs → Aim at Series A to growth-stage startups (PM hiring +2% vs 2019) and non-tech companies building AI → LLM fluency is the baseline now Designers → The hardest market: Big Tech design hiring is −48% vs 2019, startups −22% → Referrals over postings, and close the build gap New grads → The internship is the hiring event. Each one now gets ~273 applications (Handshake) → Non-tech employers are the bigger pool
1
67
11/ That's the thread. The full breakdown is on my Substack: → All 17 charts → Every source linked: 10-Ks, BLS, Challenger, SignalFire, Indeed, Levels.fyi, CBRE, Handshake, Gem → The full playbook by role, plus what would change my mind jugaldb.substack.com/ Huge thanks to @lennysan. His job market report is the best regular read on this market, and this deep dive started with one line in it. If this was useful, repost the first post so it reaches someone who's job hunting. Follow @jugaldb for more data-backed takes on tech careers and AI. Views my own, not my employer's.
39
Just got access to @claudeai duet, what are people using it for?
38
Just asked my @claudeai agents to finish work before GTA VI is playable. It’s gonna be either them finishing or me canceling my subscription
1
2
124