m
mahabubealah606

Mahabube A.

@mahabubealah606

AI Agent, RAG and Voice AI Engineer

Bangladesh
Anglais, Bengali
Certaines informations sont présentées en anglais.
À propos de moi
I build production-ready AI agents, RAG chatbots, voice assistants, and Python/FastAPI integrations. My work includes LangChain tool calling, document search with Qdrant and PGVector, real-time STT-LLM-TTS voice pipelines, REST and WebSocket APIs, PostgreSQL, Docker, and AWS deployment. I focus on reliable architecture, clean code, measurable evaluation, and clear communication from prototype to production. Whether you need an AI chatbot trained on your documents, an automated agent connected to your APIs, or a multilingual voice solution, I can design, build, test, and deploy it.... Plus d’infos

Compétences

m
mahabubealah606
Mahabube A.
hors ligne • 
Temps de réponse moyen de 1 heure

Voir mes services

Intégrations IA
I will build a rag chatbot trained on your business documents
De90 $US / projetPlus de détails
Intégrations IA
I will build custom business ai voice agents with vapi, retell, and n8n automation
De120 $US / projetPlus de détails

Expérience professionnelle

Empower_Your Business with Cutting-Edge AI Solutions

AI Engineer

Empower Your Business with Cutting-Edge AI Solutions • Temps plein

Jul 2025 - Sep 2026 • 1 yr 2 mos

Jawib - AI Agent, Voice, and Automation Platform • Built a multi-tenant platform for AI assistants across web chat, browser voice, and phone calls, with agent memory, guardrails, tool calling, and configurable conversation flows. • Developed real-time voice pipelines with STT, LLM, TTS, VAD, streaming audio, interruption handling, and SIP/PBX integrations for inbound and outbound calls. • Built visual workflow automation with LLM, HTTP, condition, transform, wait, parallel, and sub-flow nodes, plus webhook and scheduled execution. • Developed public REST APIs, JavaScript and Python SDKs, embeddable widgets, and integrations using OAuth, API keys, and custom authentication headers. • Fine-tuned Arabic speech recognition models and integrated speech services into voice-agent prototypes and pipelines. • Optimized voice pipeline responsiveness to approximately 600 ms-1 s end-to-end latency, depending on configuration Enterprise Document RAG Platform • Built an end-to-end document RAG pipeline for Arabic and English business files, spanning OCR, layout processing, chunking, embeddings, vector search, and source-grounded answers.