Full Stack AI Developer
FastAPI | Next.js | AI Agents | Voice Systems | Data Products
Building practical AI systems with reliable APIs, clean UX, and production-minded engineering.
| Area | Engineering focus |
|---|---|
| Agent systems | Tool-calling reliability, constrained outputs, failure-safe behavior |
| Product delivery | FastAPI backend + modern React/Next.js frontend |
| Voice and realtime | Low-latency STT -> model -> TTS loops |
| Data projects | Clear storytelling with reproducible analysis pipelines |
| OSS workflow | High-velocity iteration across active upstream forks |
Selection logic: recency of updates, technical depth, and direct end-user usefulness.
| Rank | Project | Type | Updated | Why it stands out |
|---|---|---|---|---|
| 1 | omnidev | Original | 2026-02-15 | Full-stack AI platform combining DevOps, scraping, vision, and storage workflows |
| 2 | feb_challenge | Original | 2026-02-21 | End-to-end IPL analytics with large-scale dataset processing and interactive visualization |
| 3 | eklavyasubmission | Original | 2026-01-18 | Education-focused AI explainer with structured exam-style output |
| 4 | langchain-rag-tutorial-2026 | Original | 2026-01-22 | Modern RAG patterns using agents, local model support, and streaming |
| 5 | docs-agent | OSS fork work | 2026-02-22 | Kubeflow documentation assistant architecture with RAG and K8s serving stack |
All images below are stored in this repo under assets/cards/ (no external image hosting).
OmniDev: platform architecture
Core modules
- DevOps agent for AWS command execution
- Stealth scraping workflows via Playwright
- Vision/OCR assistant workflows
- S3-driven storage operations
- Location intelligence API endpoints
flowchart LR
UI["Next.js Frontend"] --> API["FastAPI API Layer"]
API --> AGENT["Agent Orchestrator"]
AGENT --> LLM["OpenAI Models"]
AGENT --> AWS["AWS APIs"]
AGENT --> PW["Playwright Worker"]
AWS --> S3["S3 Storage"]
IPL evolution analysis: data pipeline
Dataset scale
- 278,205 deliveries
- 1,169 matches
- 17 seasons (2008-2025)
Output
- 10 interactive Plotly visualizations
- Narrative findings around run-rate growth, bowling adaptation, and toss impact
flowchart LR
RAW["Cricsheet Raw Files"] --> CLEAN["Python Data Processing"]
CLEAN --> DATASETS["Structured CSV Datasets"]
DATASETS --> NB["Analysis Notebook"]
NB --> VIS["Interactive Plotly Visuals"]
VIS --> STORY["Findings and Storytelling"]
AI Concept Explainer: product loop
Product direction
- Topic input for JEE/NEET syllabus areas
- Step-by-step explanations, worked examples, and MCQs
- Math rendering and exam-oriented learning flow
sequenceDiagram
participant S as Student
participant UI as React Frontend
participant API as Edge Function
participant M as AI Model
S->>UI: Enter topic
UI->>API: Request explanation
API->>M: Prompt with exam constraints
M->>API: Structured explanation output
API->>UI: Steps + example + MCQ
UI->>S: Render formatted learning module
| Repository | Updated | Focus |
|---|---|---|
| docs-agent | 2026-02-22 | Kubeflow documentation RAG assistant |
| sdk | 2026-02-22 | Python SDK for AI workloads on Kubernetes |
| trainer | 2026-02-22 | Distributed training and LLM fine-tuning workflows |
| pipelines | 2026-02-22 | Kubeflow pipelines ecosystem work |
| gemini-cli | 2026-02-22 | Terminal-native AI agent tooling |
| jenkins | 2026-02-22 | CI/CD platform fork maintenance |
- Email: jhahimanshu653@gmail.com
- LinkedIn: linkedin.com/in/himanshu748
- GitHub: github.com/himanshu748