About / Engineering / Toolkit / AI Radar / Explore / Connect
Always curious. Always building.
Visit my portfolio ↗ · Browse my repositories ↗
I'm Manikant Kumar, a developer working across AI-augmented full stack development, backend engineering, and blockchain smart contracts. I turn complex problems into clear architectures, connected workflows, and useful software.
My current focus is agentic AI, retrieval-augmented generation, secure APIs, and scalable backend systems. I care about how the pieces work together—from the interface and data layer to authentication, deployment, and observability.
|
LLM API integrations, semantic search, RAG pipelines, vector databases, multimodal interfaces, and agents that work with tools. |
Java, Spring Boot, Python APIs, .NET, microservices, authentication, and system design with reliability and maintainability in mind. |
|
MERN, TypeScript, React, Vue, and Three.js—connecting thoughtful interfaces to the systems behind them. |
Solidity smart contracts, Ethers.js, and blockchain integrations, alongside Docker, Kubernetes, cloud platforms, and CI/CD. |
Meet ManikantChatsU · Explore my profile through an AI × Web3 assistant.
| Layer | What I focus on |
|---|---|
| Experience | Responsive interfaces, full stack applications, clear API contracts, and useful interactions. |
| Intelligence | Embeddings, retrieval, model integration, agent orchestration, and tool-driven workflows. |
| Services | Java / Spring Boot, Python / FastAPI, .NET, Node.js, microservices, and secure authentication. |
| Data | PostgreSQL, MongoDB, SQL, Redis, vector search, Pandas, and NumPy. |
| Delivery | Containers, Kubernetes, CI/CD, Linux, AWS, Azure, and learning GCP. |
| Trust | API security, DevSecOps, prompt injection defense, data leakage prevention, and smart contract security. |
How I approach a build: understand the problem → define service and data boundaries → build a useful slice → test the behavior → deploy and observe → improve.
Currently exploring — the learning never stops
- AI foundations: deep learning, reinforcement learning, neural networks, model training, and decision systems.
- Agent workflows: LangGraph, AutoGen, CrewAI, n8n, automation, and coding agents.
- Retrieval and multimodality: FAISS, Pinecone, Weaviate, semantic search, vision, speech, and image + text integration.
- Systems fundamentals: DSA, OOP, concurrency, memory, operating systems, Linux, DNS, HTTPS, and domain setup.
- ML operations: MLflow, reproducible workflows, deployment strategy, and monitoring.
| Focus | Tools & technologies |
|---|---|
| Model APIs | OpenAI · Anthropic Claude · Google Gemini · DeepSeek · Mistral · Hugging Face |
| RAG & agents | LangChain · LlamaIndex · LangGraph · AutoGen · CrewAI · n8n |
| Retrieval & ML | Pinecone · FAISS · Weaviate · PyTorch · TensorFlow · scikit-learn · MLflow |
| Web3 | Solidity · Ethers.js · MetaMask · Solana |
| Everyday engineering | Git · GitHub · GitLab · GitKraken · Docker · Kubernetes · CI/CD |
More of the stack
C, Go, HTML, CSS, Express, Three.js, NumPy, Pandas, pylab, Nginx, Bash, Jenkins, Terraform, and enterprise programming patterns. GCP is part of my current learning path.
A curated timeline of the releases shaping how I think about AI engineering.
Last checked: November 3, 2026. This is a curated snapshot; availability varies by product and access tier.
| Announced | Model / release | What changed | Access at announcement |
|---|---|---|---|
| Oct 30, 2026 | GPT-6 Rosalind Research | GPT-6 reasoning for multistep life sciences research. | Limited to approved Research customers; separate approval required. |
| Oct 28, 2026 | Gemini 4 Carbon | Focus on sustained engineering tasks and cybersecurity defense. | Initial access for selected cyber defenders and trusted partners; broader release pending. |
| Oct 26, 2026 | GPT-6.1 Terra | Visual reasoning, computer use, and writing. | Rolling out in the API, ChatGPT Work, and Codex. |
| Oct 13, 2026 | Gemini 4 Argon | Coding, agentic workflows, mathematics, and science. | Generally available through the Gemini API. |
| Oct 7, 2026 | Claude Haiku 5.5 | Small model aimed at fast, high-volume workflows. | See official release for platform availability. |
| Sep 28, 2026 | Claude Sonnet 5.5 | Faster, lower-cost complement to Opus 5.5. | See official release for platform availability. |
| Sep 22, 2026 | Claude Opus 5.5 | Agentic coding and knowledge work. | See official release for platform availability. |
My builder takeaway: evaluate models on the actual workflow—quality, latency, cost, tool reliability, and security—then design routing and fallbacks around the result.
Timeline notes & official sources
Selected milestones, rather than an exhaustive history or a benchmark ranking. The 2026 “agents + tool use” card describes an engineering direction, not a single model release. The release table is a dated snapshot and does not update automatically.
- 2020: GPT-3 paper.
- 2022: Introducing ChatGPT.
- 2023: GPT-4.
- 2024: GPT-4o · Claude 3.
- 2025: DeepSeek-R1 · GPT-5.
- 2026: OpenAI model release notes · Anthropic announcements · Gemini API changelog · Gemini 4 Carbon announcement.
Choose a topic to browse matching repositories on my GitHub.
I share ideas and learning on Medium and DEV. Find my wider project links on Linktree.
GitHub activity
Activity cards are provided by external services and may occasionally be unavailable.
Behind the code
When I'm away from code, you'll often find me reading novels. Curiosity connects both worlds.
Have an idea involving AI applications, backend systems, smart contracts, or system design? Let's talk.
More places to find me
Instagram · Facebook · Dribbble · CodeChef · Codeforces · Topcoder · Stack Overflow
