You will get I will build a production AI agent with RAG and multi-channel integration
Rising Talent
Project details
You will get a production AI agent — not a chatbot demo, not a generic ChatGPT wrapper.
I build agents that:
• Retrieve from YOUR knowledge base using RAG, with citations on every claim
• Run across web, Telegram, WhatsApp, Discord, voice — one brain, every channel
• Call tools with typed schemas, security checks, and span-level logging
• Handle multi-language (I shipped a bilingual Bangla/English legal RAG agent in production)
• Sit behind observability so you see what users ask, what works, and what to fix
Behind the work: 13+ years building software, plus production AI systems including multi-agent ops platforms, voice
agents, and RAG infrastructure on AWS, Supabase, and self-hosted stacks (Ollama, FLUX, fish-speech).
What you'll get:
• A working agent deployed to your chosen channels
• Documented system prompt, retrieval pipeline, and tool registry
• Observability dashboard so you can audit every conversation
• Source code, environment config, and a handover walkthrough
If you've outgrown button-flow bots, or want infrastructure not slideware, this is the agent build for serious use.
Scope on a call before order — I'd rather pass than over-promise.
I build agents that:
• Retrieve from YOUR knowledge base using RAG, with citations on every claim
• Run across web, Telegram, WhatsApp, Discord, voice — one brain, every channel
• Call tools with typed schemas, security checks, and span-level logging
• Handle multi-language (I shipped a bilingual Bangla/English legal RAG agent in production)
• Sit behind observability so you see what users ask, what works, and what to fix
Behind the work: 13+ years building software, plus production AI systems including multi-agent ops platforms, voice
agents, and RAG infrastructure on AWS, Supabase, and self-hosted stacks (Ollama, FLUX, fish-speech).
What you'll get:
• A working agent deployed to your chosen channels
• Documented system prompt, retrieval pipeline, and tool registry
• Observability dashboard so you can audit every conversation
• Source code, environment config, and a handover walkthrough
If you've outgrown button-flow bots, or want infrastructure not slideware, this is the agent build for serious use.
Scope on a call before order — I'd rather pass than over-promise.
AI Algorithms
Large Language Model, Multimodal Large Language Model, Transformer ModelAI Applications
AI Chatbot, AI Text-to-Image, AI Text-to-Speech, AIOps, Automatic Speech Recognition, Conversational AI, Machine Translation, Natural Language Generation, Natural Language Understanding, Speech SynthesisAI Development Language
PythonAI Tools
Gradio, Hugging Face, PyTorch, StreamlitAI Models
BERT, ChatGPT, GPT-4, LLaMA, OpenAI Codex, Stable Diffusion, WhisperWhat's included
| Service Tiers |
Starter
$249
|
Standard
$799
|
Advanced
$1,999
|
|---|---|---|---|
| Delivery Time | 7 days | 14 days | 21 days |
Number of Revisions | 2 | 3 | 5 |
AI Model Integration | |||
Batch Normalization | - | - | - |
Database Integration | - | ||
Detailed Code Comments | |||
Image Upscaling | - | - | - |
MLOps | - | - | |
Model Deployment | |||
Model Documentation | - | ||
Model Monitoring | - | - | |
Model Testing & Optimization | - | - | |
Model Tuning | - | - | |
Natural Language Processing | |||
NLP Tokenization | |||
Pre-Training | - | - | |
Prompt Engineering | |||
Setup File | |||
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$149 - $699
Additional Revision
+$45
Extra channel integration
(+ 2 Days)
+$199
Voice (TTS or full loop)
(+ 2 Days)
+$299
30-day post-launch support
+$299Frequently asked questions
About Rasel
AI Agents + Full-Stack Dev | RAG, Next.js, FastAPI, AWS | 13+ yrs
Dhaka, Bangladesh - 6:59 am local time
Two cores, one practice:
→ AI Systems — multi-agent operations, bilingual RAG, voice + multi-channel agents (web · Telegram · WhatsApp · Discord), custom AI workflows, evals + observability.
→ Web Apps — e-commerce, custom SaaS, member portals, kiosks, APIs + backends, Postgres-first databases, performance + platform.
Anchored by 13+ years running real businesses across digital and physical products — Custom Cap BD, US Custom Caps, OPM International. Every web app I ship knows the physical reality it sits on: orders, production windows, supplier capacity, shipping constraints. The dual perspective is the work.
What I'm shipping right now:
• jobXlaw — bilingual Bangladesh labour-law RAG + CV studio. Next.js + Supabase + Convex. One agent brain across web → Telegram → WhatsApp → voice.
• RedClaw + GoClaw — multi-agent gateway routing 9 specialist agents across Discord, WhatsApp, and an ops dashboard. Subscription-backed inference ($200 Claude + $100 ChatGPT) running workloads token APIs would charge $2,000+/mo for. Live, in production.
• Marina Rewards — operator intel dashboard + member portal + fuel-dock kiosks. FastAPI + React + Celery on AWS ECS Fargate + RDS + ElastiCache.
• Labor Law Partner — sole architect, Mar–May 2026. Shipped a bilingual labour-law platform end-to-end in 2 months (Next.js · Clerk · Convex · Supabase · RAG · multi-agent). Full handover at close.
Stack I lean on:
Next.js · TypeScript · Tailwind · shadcn/ui · Supabase · Convex · Postgres · pgvector · FastAPI · Python · Celery · Redis · Docker · AWS ECS · Vercel · Sanity · OpenAI · Anthropic · Grok · OpenRouter · RAG · agent orchestration · CI-integrated review agents.
How I work:
Plan first. Wire both ends — auth, data, routing, billing, email, observability. Audit logs on anything that moves money or touches user data. Ship the thin slice, iterate against real use. No half-finished implementations.
What buyers get:
End-to-end ownership. I don't hand back fragments — every engagement closes with a deployable surface, structured handover, and clean account migration if you want to take over operations. Source code, env, infrastructure ownership — yours.
If you're building something AI-native, or a web platform that needs to stop feeling like a prototype, let's talk.
Steps for completing your project
After purchasing the project, send requirements so Rasel can start the project.
Delivery time starts when Rasel receives requirements from you.
Rasel works on your project following the steps below.
Revisions may occur after the delivery date.
Discovery and architecture proposal
Review your requirements, audit the knowledge sources, and confirm channels. Deliver a brief architecture plan covering LLM choice, retrieval strategy, and integrations before any code.
Knowledge base and RAG setup
Ingest your sources, chunk and embed them, and stand up the vector database. Validate retrieval accuracy with your sample queries before agent wiring.