👋 Hallo, I'm Amit Kumar
✍️ My work spans end-to-end AI platforms, covering agent orchestration, LLMOps, cloud deployment, and real-world scalability.
✍️ I'm actively exploring Recursive Language Models (RLM) and Context Engineering to enable more adaptive, memory-aware, and reasoning-driven AI systems in Year 2026.
- 🔭 Designing and operationalizing Agentic AI, RAG pipelines, and LLMOps platforms
- 🧠 Exploring recursive reasoning, long-horizon context design, and multi-agent systems
- 🤝 Open to collaboration on GenAI platforms, AI infrastructure, and applied research
- 💬 Happy to discuss AI architecture, LLM systems, and production ML
- 💻 Open-source projects available on 👉 GitHub
- 📫 Connect with me:
- ☕ Support my work:
Fun fact: I collect books faster than I can read them — they double as inspiration and décor
| Category | Technologies |
|---|---|
| 🤖 LLM Models | GPT-4, Claude 3, Gemini, LLaMA 3, Mistral, Qwen, DeepSeek |
| 🔗 Orchestration | LangChain, LangGraph, CrewAI, AutoGen, Haystack |
| Vector DBs | Pinecone, Weaviate, Qdrant, Milvus, Chroma |
| 🔍 RAG Stack | Semantic Search, Hybrid Search, Reranking, Context Engineering |
| ⚙️ LLMOps | Prompt Engineering, Model Evaluation, Cost Optimization |
| ☁️ Deployment | Azure, GCP, Vertex AI, Docker, Kubernetes , AWS |
| Domain | Proficiency | Domain | Proficiency |
|---|---|---|---|
| 🧠 RLM (Recursive LM) | 🔗 Agentic AI Systems | ||
| 📚 RAG Architecture | ✍️ Prompt Engineering | ||
| 🧩 Context Engineering | ⚙️ LLMOps & Deployment |
🧪 Experimenting With
- DeepSeek for cost-effective inference
- Hybrid RAG + Agentic workflows
- Production LLMOps patterns
Current Mood: Claude-3-Opus 🧠
Thoughtful, detailed, and overthinking every edge case in production
Other Personas:
- 🚀 GPT-4 Mode: Fast, creative, shipping features
- 🔬 Gemini Mode: Researching, experimenting, testing
- ⚡ Groq Mode: MAXIMUM SPEED coding sessions
- 🦙 LLaMA Mode: Open-source everything!
- 🧘 Claude Mode: Carefully architecting systems
Updated daily based on coffee intake ☕
Severity Levels:
- 🔴 CRITICAL: Typo in my name
- 🟡 HIGH: Broken link
- 🟢 LOW: You don't like my color scheme
- 🔵 FEATURE REQUEST: Want more jokes