TechTokWithKriti

Idk why I never used an extension for doc highlights on my browser! My life is being made so much easier with Glasp extension on @googlechrome

Traditional cybersecurity roles don’t fully cover the unique risks of AI, such as agent exploitation, data leakage & secure orchestration of tools & APIs. This is why companies like OpenAI, Okta, TD Bank, Deloitte etc are now explicitly hiring AI focused security engineers.

Coding agents are excellent support tools. They can even build full components when the boundaries are clean and isolated. But real production systems are rarely that simple. They’re interconnected, opinionated, and full of context an agent doesn’t have.

If you are using coding agents to build and are submitting MRs even without a cursory look at what is being generated: (a) Your speed of execution will be very fast (b) Your team will start loosing trust in you, just as fast

Finished my accrued monthly ClassPass credits in 1 week Two things: - Working out post vacation is torture, and yet amazing - How are workout classes in Toronto sooo expensive?

Claude Opus 4.5 dropped, and the upgrades are real. Long-horizon reasoning got sharper, multi-step tool use is smoother, and coding reliability jumped. For engineers, this means fewer hallucinated steps and more workflows you can trust end-to-end. Have you used it yet?

Shoutout to Cloudflare for giving the whole tech industry a surprise mental-health break this week. My backlog didn’t enjoy it though #meme

Gemini 3.0 is here 🔥 The multimodal reasoning jump is real, and the agentic coding upgrades feel built for engineering workflows. As builders, we get faster debugging, better context handling, and AI that can follow multistep workflows. Excited to play with it!

TechCrunch highlighted how VCs are raising the bar for AI startups and expecting speed and infra maturity from day one. As someone who leads a platform engineering team, I can tell you this: developer velocity is the new competitive advantage

For more context on cadence: It lets devs write durable, long-running workflows in code while the engine handles all the messy distributed-systems problems behind the scenes.

Strong signal that workflow engines are becoming a formal layer of cloud native stack Cadence Workflow which was originally built at Uber (and a cousin to Temporal) just entered the CNCF sandbox.

A new role is blowing up in AI companies: Forward Deployed Engineer (FDE). You build, deploy, and debug on the frontlines with customers. It’s not research, but is closer to being an embedded engineer-consultant who makes AI work in the real world.

Meta’s recent layoffs in some AI teams and yet doubling down on others (‘super intelligence’ rhetoric) mean reorganizing more than anything else. Imo, they are choosing their battles when it comes to AI. #metalayoff #meta #MAANG

Every outage teaches the same two lessons: 1️⃣ Observability isn’t optional, we need to be able to detect anomalies early. 2️⃣ Resilience is not just the provider’s responsibility. Practice regional failure like fire drills. AWS just showed why both matter for every customer

I picked this topic because running the same app across clouds gets messy fast with multiple pipelines, config drifts. Karmada lets you keep the Kubernetes mindset and still roll out cleanly across clusters with policy. That felt worth putting on a stage.

Got to share “Deploy Once, Run Anywhere” at @AllThingsOpen 2025 and am still buzzing about it.

Sanyogeeta Lawande, at Ford Motor Company nominated for the #WomenTech Global Awards 2024! Share to raise awareness! #womenintech #womeninstem #womenwhocode #WTGA2024 @womentechnet womentech.net/nominee/all/al…

@tnslxk @X Hi Sunita Nice to meet you! I am in the Cloud domain!

If you're in Tech Let's connect #conncet #developers @X

Apple users: Please provide your reviews and suggestions?

Apple Products to purchase this year? - Need a new phone: Is the new iPhone 16 worth it? I primarily use my phone for content creation + some editing