Aleksei Serebriakov 🐍

Me to Claude: "Make no errors. DO NOT HALLUCINATE. YOU ARE AN EXPERT SOFTWARE ENGINEER "

@FlippedRay snapapi.pics

I’m trying new products today 👇 Drop yours

My first attempt at ricing MacOS, wha do you guys think? #ricing #macos

@Tobby_scraper snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

GM Guys. Last day of the week. Shipping something?

@JacobCounsell snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

Happy Friday Vibe Coders! Who’s up building right now?

@CiprianiRanieri snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

drop ur startup! I am curious

@yusukelp snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

GM builders☀️ It’s Saturday Share your saas/app/project below!!

@sickdotdev snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier

What are you building this weekend? Drop your portfolio URL Consider this as marketing.

@sickdotdev snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

@mscode07 snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

Founders, what are you building? Pitch it in 1 sentence.

@rxhit05 snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

Hey founders!! Looking to connect with people building in: 💻 SaaS ⚙️ Tech 🤖 Automation 🧠 AI tools 📦 Product Development 🌐 Web apps Drop what you're working on 👇🏼

@delveroin snapapi.pics , ultimate web screenshot/video api, with support of scraping, extraction, llm analysys and s3 vault support, 8 sdks, 200calls/mo free tier forever

What are you building this weekend? Drop your portfolio URL Let’s send some traffic

@mscode07 Building snapapi.pics, ultimate screenshot api, happy to connect :)

hey builders, let's connect and grow together drop what you're building below👇

@BuffetDesigns Building snapapi.pics, ultimate screenshot api, happy to connect :)

Looking to connect with more founders, devs and designers who are building! Drop a comment if that's you👇

@foundrceo Hey, happy to connect, i'm building snapapi.pics, ultimate screenshot api :)

Looking to connect with more founders, devs and designers who are building! Drop a comment if that's you👇

@NaivaidyaY66600 snapapi.pics screenshot api with support of videos, data extraction, llm analysys, s3 vault, proxies etc

Hey Founders, what are you building right now? Drop your links 💪💪

@heyblake snapapi.pics screenshot api with support of videos, data extraction, llm analysys, s3 vault, proxies etc

Fork it Drop your landing page URL I'll give 1 piece of advice to as many of you as I can

KEDA changes the conversation — scaling on Kafka lag, SQS depth, or Redis list length is way more useful than CPU for any async workload. The scaleTargetRef + minReplicaCount: 0 combo lets you drop worker pods to zero during idle, which single-handedly cuts the bill on BullMQ-style queues.

Day 12: Kubernetes Autoscaling Design. "How does your k8s cluster scale?" → Most candidates say HPA. Interviewers want all 3 layers. Layer 1: HPA (Horizontal Pod Autoscaler) → Scale pod replicas based on CPU/memory or custom metrics → Custom metrics: Prometheus Adapter → KEDA (event-driven) → Stabilization window matters — prevent flapping Layer 2: VPA (Vertical Pod Autoscaler) → Right-size pod resource requests/limits automatically → Don't run VPA + HPA on same metric simultaneously Layer 3: Cluster Autoscaler / Karpenter → Add/remove nodes based on pending pods → Karpenter (AWS) → provisions nodes in seconds vs minutes → Spot instance handling: interruption detection + graceful drain Interview signal: "We use KEDA for queue-based scaling — scale to zero when queues are empty." Cost optimization + scale-to-zero = senior flag.

For "charge a card" the answer most candidates miss is the ledger is the source of truth, not the payment gateway — append-only entries with double-entry accounting invariants, then reconcile the gateway state against it via a scheduled sweep job. That framing also solves the webhook out-of-order problem.

14/ Company‑specific: Razorpay / Stripe / fintech Key themes: - Consistency, correctness, idempotency - Observability: logs, metrics, alerts - Security mindset Prep: - Know exactly how you’d design “charge a card”, “retry a failed payment”, “webhook with retries + backoff.” - Have real examples of bugs you prevented / debugged in critical flows.

1/ Most people “prepare for interviews”. If you’re aiming for top spots like @Rippling , @databricks , @stripe , @nvidia , @Razorpay , @JioHotstar etc, you may need to rebuild yourself for that bar. Here’s the PLAYBOOK for 0–5 YOE folks in India. Save this. 🔖

Unique index on event.id + INSERT first, then catch 23505 and 200 → gives you dedup for free. For out-of-order the trick is reading subscription.status from a Retrieve API call inside the handler rather than trusting the event payload — Stripe has eventual consistency on the object but the API is canonical.

@kayintveen @fidexcode stripe webhooks specifically yeah. spent more time on webhook idempotency than the actual checkout flow. events show up out of order, duplicates fire randomly — claude gives you a clean handler that works great until real traffic hits it

Forget hype Is it possible to build an e-commerce website in a week with AI?