Aman

1.9K posts

Aman banner
Aman

Aman

@amandotdev

swe & ai enginner • 2x Internship • oss • Available for freelancing

Tham gia Mayıs 2022
399 Đang theo dõi206 Người theo dõi
Aman
Aman@amandotdev·
i was woking on distributed job scheduler. thought to make some notes about the distributed systems and project structure.
Aman tweet media
English
0
0
3
21
Aman
Aman@amandotdev·
@16_shivang How ur netmirror is working Its showing cloudflare error for me 😭
English
1
0
0
282
Devanshu
Devanshu@DevanshuXi·
Had a really interesting @helixdb interview recently. One of the rounds involved designing a KOPS-like cluster orchestration system from scratch, covering AWS provisioning, custom control plane design, failover, split-brain prevention, leader election, and distributed health checks. The discussion got especially interesting when we started modeling cluster topology as a graph. Each node represented a machine or service, while edges captured dependencies, replication links, and communication paths. From there, we talked about optimizing graph traversal using DFS while maintaining node state in a hashmap. This makes cycle detection trivial, which is critical for identifying replication loops, dependency deadlocks, or invalid failover chains. A hashmap also helps cache intermediate results, avoiding repeated traversals when evaluating large cluster topologies. That becomes incredibly valuable when the graph spans thousands of nodes across multiple regions. It was one of those interviews where distributed systems, graph theory, and practical infrastructure all came together beautifully. Learned a ton from this one. What a quality.
Devanshu@DevanshuXi

Hey! I’m actively looking for a full-time backend / infrastructure engineer role (new grad 2026). I’m pretty tech-stack agnostic, but I naturally gravitate towards Go and problems around distributed systems, databases, and performance. I like building systems that actually move data at scale and understanding why they behave the way they do under load. I interned at Cal.com, working on scheduling infrastructure, where I handled complex challenges that come with time. I’ve also been contributing to Vitess, working on VTGate internals, specifically around query routing, session handling, and correctness in distributed MySQL clusters. Outside of that, I spend a lot of time building systems to push my own understanding: 1. Built a distributed search engine with sharding + fan-out query execution + Top-K merging, hitting ~7K QPS at ~28ms p99 2. Designed a hybrid ranking pipeline (BM25 + embeddings) to handle both lexical and semantic search 3. Built a git-like schema versioning system with a commit DAG, merge logic, and deterministic migrations for zero-downtime environments I also write about what I build and learn: @devanshusharma658" target="_blank" rel="nofollow noopener">medium.com/@devanshusharm… Mostly deep dives into distributed systems, query execution, and performance tradeoffs. I care a lot about performance (p99s, not just averages), clean abstractions, and observability, and I try to build systems that are boring and reliable in production. If you're hiring for backend/infra / distributed systems roles (remote or relocation-friendly), I’d love to chat. And honestly, if your team is working on hard, complex problems, I’d love to dive in. Note: I’m also learning Rust these days because, eventually, giving less pain to your CPU is one of my goals.

English
5
2
153
9.4K
Aman
Aman@amandotdev·
I want to go in the ML infra side and applied ai like scaling application/ inference engineering and along with distributed systems. Code writing is easy nowdays but if somebody has some guidance from where i can start please let me know .
English
1
0
4
64
Aman
Aman@amandotdev·
@dims What type of projects i need to make for this role 💀😭 Exp is kinda tough to get
English
0
0
0
63
Aman
Aman@amandotdev·
@0xSero Bro i like this what u r posting last day about quantisation and now inference this is getting me alot of knowledge pls continue to post more
English
0
0
1
253
0xSero
0xSero@0xSero·
Here's everything you need to know about inference and hosting LLMs. Have you ever seen: - vllm - sglang - llama.cpp - exllamav3 these are all engines that allow us to run LLMs, it's not easy but if we work together it will be. youtube.com/watch?v=B18zBn…
YouTube video
YouTube
English
5
66
582
31.8K
Ashikka
Ashikka@AshikkaG·
2nd place at @opencode Buildathon 🥈 @BansalRishit and I built openflip.io - a pocket AI red teaming agent that scans nearby signals (Sub-1GHz, Bluetooth, WiFi, NFC, RFID, IR), finds/generate the right exploit modules via @opencode, and runs them in real time. Flipper Zero… but it thinks for itself. 👉 Join the waitlist: openflip.io Huge shoutout to @GrowthX_Club, @opencode, @udayan_w, @nexxel and the entire team for hosting such an incredible event. Second win this weekend - we are working on building and shipping both these projects in the coming days! Stay tuned.💪
Ashikka tweet mediaAshikka tweet media
English
58
28
1.2K
56.3K
Aman
Aman@amandotdev·
going to implement more things on this project and learn more about distributed system
English
0
0
0
19
Aman
Aman@amandotdev·
i m curently building distributed web application firewall in golang. > made a custom DSL rule engine (lexer → parser → AST → evaluator) > hot-reload via fsnotify + atomic pointer swap > token bucket rate limiter per ip/route
Aman tweet media
English
1
0
3
109
Aman
Aman@amandotdev·
@ItsRoboki Isn’t it rtk clone? How is it better
English
1
0
1
97
Jagrit
Jagrit@ItsRoboki·
Your Claude Code is burning through thousands of tokens on noise This cli tool solves it ztk sits between Claude Code and your shell and compresses the output before it hits the model Same info, way fewer tokens Been running it for a while now and it reduced my context usage drastically
Jagrit tweet media
English
7
2
54
4.4K
Ajay Bhakar
Ajay Bhakar@ajay_2512x·
I am mostly applying to AI engineer roles or Backend roles with AI architecture focus. So far has 1 interview last week , really strong tech round but not positive about the output because of discussions related to camera access. Please let me know what and where am I making mistake and what can I improve what would you advice him?
Ajay Bhakar tweet media
English
5
4
72
7.6K
Aman
Aman@amandotdev·
@Hi_Mrinal Bro do i write code by ai or hand😅 I architect most of my things like how it works But for writing today ai is becoming more prominent
English
1
0
1
152
Aman
Aman@amandotdev·
@kmeanskaran @UnslothAI Even if i learn this then how can i make projects based on this Like which can schedule reflect my skills
English
0
0
1
177
Karan🧋
Karan🧋@kmeanskaran·
These are actual high-demand next-level AI skills, not RAG: - Data prep for instruction fine-tuning - @UnslothAI ecosystem for fine-tuning, reasoning models, quantization, etc. - Fine-tuning embedding models - Backend design using FastAPI, Redis caching, queue workers, rate limiting - Making LLM inference layer using @vllm_project - @DeepSpeedAI for multi-GPU training - Learn NVIDIA Triton for running any vision and object detection models along with LLMs - Learn about Docker and CUDA setup - Use of @huggingface ecosystem - Distributed systems using @anyscalecompute - Kubernetes and Terraform setup for serving your model - Learn about @modal for low-config setup deployment - Deploying LLM on Ollama for easier calls - Context engineering - Agent memory harness - Multi-agent orchestration - LLM as a judge These skills are more focused towards inference engineering to ensure that your model runs smoothly in production. Memory, inference, and orchestration matter a lot more than frontier models. I am learning inference engineering and agent harness nowdays. Totally worth exploring!
English
7
46
406
12.6K
Aman
Aman@amandotdev·
@Claude_Memory @github I just added claude mem but whats the best and effective way to use it without context bloating
English
0
0
0
66
Claude-Mem
Claude-Mem@Claude_Memory·
Claude-Mem is now the #1 most starred dedicated AI memory tool on @github
Claude-Mem tweet media
English
4
2
16
1.1K
Aman đã retweet
Z.ai
Z.ai@Zai_org·
@Maaztwts built OSAP to stop the constant tab-switching between Slack, GitHub, and Notion. It uses GLM-5.1 as a reasoning layer to handle multi-step tasks across different apps, with persistent memory (HydraDB) to keep track of your specific workflow habits. osap.maazx.dev
English
91
106
366
43.2K
Maaz (agentic arc)
Maaz (agentic arc)@Maaztwts·
@llmdevguy tbh i made this project with @opencode glm and minimax models. that too with 5$ and the project is got featured in z ai official twitter page :) x.com/i/status/20429…
Z.ai@Zai_org

@Maaztwts built OSAP to stop the constant tab-switching between Slack, GitHub, and Notion. It uses GLM-5.1 as a reasoning layer to handle multi-step tasks across different apps, with persistent memory (HydraDB) to keep track of your specific workflow habits. osap.maazx.dev

English
3
0
15
3.7K
Mateusz Mirkowski
Mateusz Mirkowski@llmdevguy·
OK that was fast.. GLM 5.1 was the best model in terms of quality and price. For just few days.🙈 For this price it's still good option, but instead of 72$ I would rather pay 100$ for codex. More reliable models. For GLM go for OpenCode GO for 5 usd. 4400 requests per month is not bad to play with it. It's slow, but works. If you like it go with lite. King of value stays with MiniMax 2.7.
Mateusz Mirkowski tweet media
English
49
7
259
251.8K
ShaRPeyE
ShaRPeyE@sharpeye_wnl·
reading about chunking strategies and embeddings for llm applications in class very fun, do give it a read
ShaRPeyE tweet media
English
10
0
135
5.4K