Build with Akshit

Computer Science Everynight šŸŒ™

Hi, I’m Akshit. I’m doing okay in computer science—unlike some other skills where I still struggle quite a bit.
I’m a bit camera-shy, but I’ve started putting out videos anyway.

I enjoy building projects with Python, Go, JS/TS, and Flutter for mobile. I picked up AI/ML to keep up with the world’s needs, and to prepare for corporate jobs I also learned Cloud, System Design, DSA, and LLD.

You’ll find my videos covering all of these topics.


Build with Akshit

Hey Guys, we have launched Bifrost on Product Hunt - Its World's Fastest AI Gateway (Open Source), beating LiteLLM!
Help us become Product of the day šŸ™šŸ» Upvote here - getmax.im/ak-yt-bifrost

What is Bifrost?

It’s a super fast, fully open-source LLM gateway built for scale. Written in Go with A+ code quality. Takes <30s to set up and supports 1000+ models across providers via a single API.

Key features:

- Blazing fast: 11μs overhead @ 5K RPS
- Robust key management: Rotate and route API keys with weighted distribution
- Plugin-first architecture: Add custom plugins easily, no callback hell
- MCP integration: Supports Model Context Protocol for tool orchestration
- Maxim integration: Seamlessly connects with Maxim for full agent lifecycle management, evals and observability.
- Governance: Manage budgets and rate limits across mutliple teams.

We built this because most LLM gateways couldn’t keep up with our needs at scale. We were running intensive evals and agent workflows inside Maxim, and hit real bottlenecks. Turns out other teams were facing the same.
If you’re looking for a faster, cleaner alternative to LiteLLM or similar tools, would love your thoughts.

Support on our Product Hunt page would go a long way for us! :")

1 month ago | [YT] | 7

Build with Akshit

Built an agent on n8n and also simulated it, check this video - https://youtu.be/05bek5NORPM?feature=...

4 months ago | [YT] | 4

Build with Akshit

If you think you are currently not a PRO developer then this is the year when you have to decide -

1. Either become a PRO dev and come into the creamy top layer of Software Engineers. Pick 1 domain and beat 99% folks in that domain.
2. Become a generalist and become a jack of all trades, gain some level of expertise on any tool / protocol / framework you see, code or no code. Become a Technically sound Product Owner.

Don't stay confused, take action and decision today.

5 months ago | [YT] | 71

Build with Akshit

Don't HARDCODE your prompts while building AI Agents, manage your Prompts on Cloud. Check this tutorial - https://youtu.be/rEDNZCCaW7E

5 months ago | [YT] | 8

Build with Akshit

Uploaded a new video on how I made Claude Desktop directly write content in my Notion via MCP šŸš€
Check the video here šŸ‘‡

#mcp #ai

6 months ago | [YT] | 9

Build with Akshit

How to compare your Prompts for lower latency, optimised cost and best outputs - https://youtu.be/ms-0zXuo0ks?feature=...

#ai #agents

6 months ago | [YT] | 12

Build with Akshit

If you are in AI, you must know what MCP is. The Model Context Protocol. Here is the video on it - https://youtu.be/8uXuPdPFulI?feature=...

#ai

6 months ago | [YT] | 6

Build with Akshit

How would I make an app today after building & breaking stuff for 6 yrs VS me in my first year trying to make my first app 🤔

6 years ago, I set out to build an app for my startup. Spoiler: I broke more things than I built. Today, I’d approach it differently, after working with 20+ Founders as Freelancer, Mobile Team Lead, Founding Engineer..

1. Mentality

Then: I need an app because startups are cool and 9-5s are the enemy. It was less about purpose, more about escape.

Now: What problem am I solving?' It’s not about the app—it’s about the why. Purpose drives the code, not the other way around.

2. First Step

Then: Googled 'how to make an app,' downloaded 17 IDEs, saved 100 tutorials. Procrastinated. Bored.

Now: Sketch the problem first. Talk to users. Prototype with intent. Code comes later—tools don’t define the start, clarity does.

3. Failure

Then: One crash = I’m a fraud. I’d abandon the project for weeks, haunted by imposter syndrome.
Now: Failure’s data. A buggy API call teaches me rate-limiting. A UI glitch shows me flexbox quirks. Six years of flops built a debug-first mindset.


4. Success

Then: Success = downloads. If it’s not on the App Store trending list, it’s trash.

Now: Success = impact. Did it solve the problem? Did someone’s day get better? Metrics matter, but meaning lasts.

5. Tech Stack

Then: Threw every shiny tool at the wall—Java for Android, Swift for iOS, some janky PHP backend. Result?

Now: Keep it lean. Build an MVP, get PMF & then double down on engineering. Maybe Flutter for speed-to-market, Supabase for backend-as-a-service, and Scalability beats buzzwords.

6. Learning Curve

Then: Watched a 12-hour YouTube tutorial, copied code I didn’t wrote, and prayed it worked. Spoiler: it didn’t.

Now: Master the fundamentals—data structures, async programming, state management. Master one framework deeply (say, how React’s Virtual DOM ticks) before chasing the next hot thing.

7. Debugging

Then: A null pointer exception was a personal enemy. I’d eyeball 500 lines of spaghetti code, change random semicolons, and cry.

Now: Logs are my friend—console.log() everything. Unit tests catch dumb mistakes. Systems beat panic.

Hmm, it's okay if you are starting today and you are unable to understand what I am saying, give it a go and create your own learning journey.

#apps

-------------------------
Credits - The below picture is AI generated (FLUX) & the content grammar was checked & improved by grok3.

7 months ago | [YT] | 60

Build with Akshit

WOULD YOU RATHER WATCH

7 months ago | [YT] | 16

Build with Akshit

GenAI / AI Agents Roadmap & Resources video coming today/tomorrow šŸš€šŸ˜ŽšŸš€šŸ˜Ž

7 months ago | [YT] | 16