AI Full Stack Developer ยท Open to Work

Building the next generation of intelligent <>

AI Full Stack Developer who ships production apps powered by RAG pipelines, multi-agent systems, and LLM orchestration. Comfortable from Postgres + pgvector up to a React frontend with streaming responses. Currently building autonomous tools.

SYS.LOC: 17.3850 N, 78.4867 E GPU.TEMP: 42ยฐC // IDLE AKASH_v2.6
๐Ÿง  model state Training / An autonomous research agent
0 Projects shipped
0 Years coding
0 LLM integrations
0 Tokens / day

01 // terminal_demo.sh

~/projects/aurora-rag โ€” zsh

02 // production_builds

{ "sensor_id": "iot_node_04", "voltage": 230, "amps": 4.8 }
โšก
02 live

IoT Electricity Monitor

Real-time power consumption tracking via IoT sensors + cloud storage across multiple endpoints.

IoT Sensors Cloud Storage Data Viz Git
Inspect codebase
[ALERT] SSH honeypot: Root login from 185.190.140.23 STATUS: BLOCKED
๐Ÿ›ก๏ธ
03 live

Cloud Honeypot IDS

Honeypot-based intrusion detection on cloud infra with attacker pattern analysis and threat intelligence.

Python Networking Cloud Security Linux
Inspect codebase
DELETE FROM availability_slots WHERE slot_datetime < NOW();
๐Ÿฅ
04 live

Hospital Slot Scheduler

Dual-role Doctor/Patient booking with Razorpay gateway and auto-cleanup of expired slots.

Python Flask MySQL Razorpay
Inspect codebase
hashed_pw = bcrypt.generate_hash(pw) session['user_id'] = user.id
๐Ÿ“
05 beta

Secure Notes Manager

Flask CRUD app with OTP-based 2FA, bcrypt auth, and per-user data isolation at query level.

Python Flask MySQL Bootstrap
Inspect codebase

03 // neural_profile

Half builder, half scientist.

I design and implement autonomous pipelines that think in loops. My code bridges the gap between raw research and high-performance, real-time user experiences.

Obsessed with context window limits, vector embedding clusters, and reducing time-to-first-token. I build systems that don't just respond, but reason.

Currently working at the intersection of developer tooling and cognitive frameworks, developing agent workflows that interact seamlessly with modern databases.

EPOCH: 48 // VAL_LOSS: 0.124 STEPS x1000
  • Name Akash Belide
  • Primary Role AI Full Stack Developer
  • Location Hyderabad, India
  • Academic Focus CS & Cognitive Systems
  • Core Interest Multi-Agent Orchs & RAG
  • Availability Open to Work

04 // system_dependency_matrix

Languages

Python
TypeScript
JavaScript
SQL
Bash

AI / ML

OpenAI API
Anthropic Claude
LangChain
LangGraph
Hugging Face
PyTorch
LlamaIndex

Frontend

Next.js 14
React 19
Tailwind CSS
shadcn/ui
Framer Motion
Vercel AI SDK

Backend / Infra

FastAPI
Postgres / pgvector
Pinecone
Redis
Docker
AWS

05 // career_logs

2024 โ€” Present NeuralFlow AI

AI Engineer

Pioneered multi-agent task runners using LangGraph, reducing graph execution routing logic latency by 35%. Developed custom context routing scripts feeding clean document tokens to Claude models. Designed dynamic chunking indexing logic utilizing FastAPI and pgvector to scale cognitive workflows across multi-tenant environments.

2022 โ€” 2024 VectorBase Corp

Full Stack Engineer

Integrated real-time streaming chatbot user interfaces with React and Vercel AI SDK. Architected low-latency caching models for active chat states on top of Redis. Engineered robust subscription checkouts using Stripe, managing vector space utilization limits inside Pinecone to decrease cost profiles.

2021 โ€” 2022 Cognitive AI Labs

Research Intern

Finetuned base transformers (Llama, Mistral) using QLoRA adapters on narrow clinical domain tasks. Standardized data pipelines converting unstructured logs into structured training sets. Logged hyperparameter sweeps using weights & biases to minimize learning rate anomalies.

06 // connection_handshake

Got a problem worth solving?

Akashbelide98@gmail.com