Parcourir les catégories
Explorer
Fiverr Pro
Français
$
USD
Abdul Mannan
@devmannan
Level 1
Expert Agentic AI, Langchain, GenAI, N8N, RAG, LLM, Full Stack Engineer
Pakistan
Espagnol, Anglais, Chinois, Japonais, Allemand, Français, Néerlandais, Italien, Suédois, Arabe, Norvégien, Danois
À propos de moi
I am a Senior Software Engineer with 9+ years of experience in Web App Development and cutting-edge AI/LLM Integrations.
My expertise covers full stack AI development where I design, build, and deploy scalable applications while integrating advanced AI features into existing platforms.
I specialize in custom chatbots (RAG & Agentic RAGs using LangChain, LangGraph, LangFlow, LangSmith) and workflow automations with tools like n8n, Make.com, and Monday.com to simplify and optimize business processes.
Compétences
Abdul Mannan
hors ligne •
Temps de réponse moyen de 1 heure
Voir mes services
Sites web IA & Logiciel
I will develop ai saas app ai chatbot ai web app ai agent ai application developer
Automatisations et agents
I will be generative ai ml llm ai agent ai chatbot python and full stack developer
Portfolio
Expérience professionnelle
.png){kind=logo}
AI Lead / Full Stack Engineer
Fiverr • Indépendant
Feb 2019 - Present • 7 yrs 6 mos
Full-Stack Engineer, Solution Architect, and AI/ML Specialist with extensive experience designing and delivering scalable web platforms, cloud-native applications, and intelligent AI solutions. Full-Stack Development: React, Next.js, Node.js, Python, FastAPI, PostgreSQL, MongoDB, Microservices, REST/GraphQL APIs, Event-Driven Architectures, Authentication (JWT/OAuth), Payment Integrations, Docker, Kubernetes, AWS, GCP, and Vercel. AI & Generative AI: LLM Integration (OpenAI, Claude, Gemini), RAG Systems, LangChain, LangGraph, AI Agents, Chatbots, Vector Databases (pgvector, Pinecone, Weaviate), Prompt Engineering, NLP, Computer Vision, Embeddings, and MLOps. Successfully delivered 50+ products including SaaS platforms, marketplaces, enterprise dashboards, AI assistants, e-commerce solutions, booking systems, recommendation engines, and knowledge-base chatbots for clients across North America, Europe, Asia, and Latin America. Led and mentored engineering teams of 3–30 developers, managing end-to-end product delivery from architecture and strategy to deployment and scale. Core Technologies: React, Next.js, Node.js, Python, FastAPI, LangChain, LangGraph, OpenAI API, Claude API, PostgreSQL, pgvector, Microservices, AWS.
22 Avis
| (22) | ||
| (0) | ||
| (0) | ||
| (0) | ||
| (0) |
Détails de la notation
- Niveau de communication avec le freelance
- Qualité de la livraison
- Valeur de la livraison
1 à 5 avis sur 22
Trier par
Les plus pertinents
willardadk
Client récurrent
États-Unis
Great Work, Technical feasibility plan was very useful and clearly explained the full talking avatar workflow, including LLM options, AI voice/TTS, lip-sync tools, computer vision role, APIs, risks, and MVP scope.
Utile?
daviskyle1
Client récurrent
États-Unis
Really satisfied with the delivery. Abdul added a clean multi-provider AI integration layer for OpenAI, Claude, Perplexity, Gemini, and Grok, with reusable backend services, .env API key setup, provider switching, error handling, and fallback logic. The demo features for AI product descriptions, smart...
Utile?
lucasmason818
Client récurrent
France
Very satisfied with the work. The LLM comparison was clear, well formatted, and covered GPT, Claude, Gemini, Llama, Mistral, and Qwen in a practical way. The recommendation was useful for deciding the right model for my AI feature, and the delivery was quick with good communication.
Utile?
anitakupka
Canada
The prompt review was really helpful and easy to understand. The rewritten prompts sounded much clearer and more natural, with useful suggestions for improving AI response quality and choosing the right model for my use case
Utile?
chaichan25
Client récurrent
Royaume-Uni
Really good work. The single-file Python AI news agent included RSS fetching, news categorization, OpenAI summaries, FastAPI dashboard, API endpoints, and clean error handling.
Utile?
