Available for projects

ASIK
M.

AI Voice Agent & Automation Engineer

I build AI voice and text companions that actually remember users — using LiveKit, Mem0, and Python.

My current project is a live, running AI companion handling real-time voice and text interactions. I specialize in fast delivery: voice agents, API pipelines, n8n automation, and custom AI workflows.

Working prototypes in 48 hours.

48H
Prototype Delivery
100%
Voice-First Focus
4+
AI Frameworks
1×
Live Agent Running Now

// 01 — What I Build

AI THAT
ACTUALLY WORKS

No demo-ware. No bloated frameworks. Working integrations delivered fast, built to production standards.

🎙️
VOICE AGENTS

Real-time voice AI companions with persistent memory. Built on LiveKit + Mem0 — they remember users across sessions and respond in milliseconds.

LiveKit · Mem0 · Python
N8N AUTOMATION

Multi-step workflow automation wired to your stack. AI-powered triggers, conditionals, and data flows that eliminate manual work entirely.

n8n · Webhooks · APIs
🔗
API PIPELINES

Custom API integration layers connecting your tools, AI models, and databases into coherent, observable pipelines that don't break.

REST · GraphQL · Python
🧠
AI WORKFLOWS

LLM-powered workflows with memory, tool use, and RAG. Agents that don't just answer — they act, recall context, and improve over time.

LangChain · OpenAI · RAG

// 02 — Tech Stack

TOOLS I
ACTUALLY USE

LiveKit Mem0 Python n8n LangChain OpenAI API FastAPI WebSockets Supabase Whisper TTS / STT Docker LiveKit Mem0 Python n8n LangChain OpenAI API FastAPI WebSockets Supabase Whisper TTS / STT Docker
Voice & Real-Time
  • LiveKit (WebRTC rooms)
  • OpenAI Whisper (STT)
  • ElevenLabs / TTS
  • WebSockets
AI & Memory
  • Mem0 (persistent memory)
  • LangChain / LangGraph
  • OpenAI GPT-4o
  • RAG pipelines
Automation
  • n8n workflows
  • REST & GraphQL APIs
  • Webhook orchestration
  • Zapier alternatives
Backend & Deploy
  • Python / FastAPI
  • Supabase / PostgreSQL
  • Docker / Render
  • Vercel / Netlify

// 03 — Signature Project

THIS IS LIVE
RIGHT NOW

Not a mockup. Not a demo script. A real, running AI companion in production.

System Online
PERSISTENT AI COMPANION

A voice + text AI companion that handles real-time conversations and actually remembers who you are between sessions. Built from scratch with LiveKit for sub-100ms voice latency and Mem0 for cross-session user memory.

Ask it something today. Come back tomorrow. It will recall your name, your context, your preferences. That's the differentiator.

LiveKit Mem0 Python WebRTC GPT-4o Whisper
agent.py — running
$ python agent.py --start
[INFO] LiveKit room connecting...
[OK] Room joined: asik-companion-01
[INFO] Loading Mem0 memory store...
[OK] Memory store ready — 42 users loaded
[INFO] Whisper STT pipeline active
[OK] Agent listening on port 8080
[USER] "Hey, do you remember what we talked about last time?"
[AGENT] Retrieving memory for user_id: u_82fa...
[AGENT] "Yes — you mentioned your project deadline is Friday and you prefer concise answers."
_

// 04 — How It Works

PROTOTYPE
IN 48 HOURS

No endless scoping calls. No vague roadmaps. Here's exactly what happens when you hire me.

01
Discovery Call

30 minutes. You tell me what you need — voice agent, automation, API integration. I ask the right questions, not the generic ones.

Day 0 · 30 min
02
Build & Ship

I build. No status theater. You get a working prototype link in 48 hours — live, testable, functional. Not a Figma file.

Day 1–2 · Async build
03
Iterate Fast

Feedback round. I refine, extend, and hand off clean code with documentation. You own it completely.

Day 3–5 · Refinement

// 05 — Why Asik M.

NOT JUST
ANOTHER DEV

Differentiator 01
MEMORY-FIRST BY DEFAULT

I don't build stateless chatbots. Every voice agent I ship uses Mem0 for persistent user memory — your AI actually learns and recalls. That's not a feature, it's the foundation.

Differentiator 02
REAL-TIME VOICE, NOT DEMO VOICE

LiveKit-based voice pipelines with sub-100ms latency. Not a pre-recorded demo. I've built it, it's running, and I know where the hard parts are.

Differentiator 03
FAST DOESN'T MEAN SLOPPY

48 hours is possible because I've already solved the boilerplate. You get speed AND clean, documented, maintainable code you can extend long after the engagement ends.

Differentiator 04
FULL PIPELINE, NOT JUST LLM CALLS

Voice, memory, automation, API orchestration — I handle the whole stack. No hand-offs, no gaps. From STT → agent logic → TTS → workflow automation, it's one coherent system.


LET'S BUILD
SOMETHING
THAT WORKS

Available for freelance contracts · Fast turnaround · Honest communication