Explore highlights of my work, showcasing innovative solutions in various domains.
Data driven analysis
AI workflow
Core projects
From High-Fidelity Data to Actionable Insights
Iām focusing on how high-fidelity data in material testing, simulations, telematics, or ADAS can transform into engaging dashboards. These dashboards highlight new chart types like map-based views and interactive timelines, revealing insights that unmask bottlenecks, vital paths, and patterns in complex data sets.
Use Case: Automated Data Extraction and Interactive Visualization
flowchart TD
%% Define Styles
classDef process fill:#e1f5fe,stroke:#01579b,stroke-width:2px, font-weight:bold
classDef limitation fill:#ffebee,stroke:#c62828,stroke-width:2px, font-style:italic
classDef enhancement fill:#fff3e0,stroke:#e65100,stroke-width:2px, font-style:italic
classDef highlight fill:#ffe0b2,stroke:#ff9800,stroke-width:2px, font-weight:bold
%% CFD Analysis Subgraph
subgraph CFD["ā” Large Data Set"]
A["High Fidelity Model"] --> B["Typical Excel Charts \n Stress/Temperature Maps "]
end
%% Smart Processing Subgraph
subgraph AUTO["š Smart Processing"]
C["Smart Data Extraction"] --> D["Sankey View"]
C --> E["Network Graph"]
D --> F["Lumped Model"]
E --> F["Lumped Model"]
end
%% System Model Subgraph
subgraph SIM["š System Model"]
G["Simscape Model"] --> H["Auto-Routed System"]
end
%% Interconnections with Edge Label Closer to C
A -->|Scripts| C
B --> L1["ā **Limitations:**\n- No Critical Path Identification\n- No Input Integrity Checks"]
C --> E1["š” **Benefits of Smart Processing:**\n- Input Verification\n- System-Level Analysis\n- Performance Insights\n- Critical Path"]
F --> G
%% Class Assignments
class A,B,C,D,E,F,G,H process
class L1 limitation
class E1 enhancement
%% Optional: Styling the Label Closer to C
linkStyle 0 label-position right
flowchart TD
%% Define Styles
classDef process fill:#f8f0fc,stroke:#6b46c1,stroke-width:2px
classDef data fill:#e6fffa,stroke:#047481,stroke-width:2px
classDef output fill:#fefcbf,stroke:#975a16,stroke-width:2px
%% Frontend Subgraph
subgraph UI["šØ Frontend"]
A["Drag-Drop Interface"] --> B["React Flow Canvas"]
B --> C["Workflow Visualizer"]
end
%% Backend Subgraph
subgraph API["āļø Backend"]
D["FastAPI Server"] --> E["LLM Router"]
E --> F["Custom Agents"]
D --> J["Additional FastAPI for Database Retrieval"]
end
%% Knowledge Base Subgraph
subgraph RAG["š Knowledge Base"]
G["Document Store"] --> H["Titan v2 Embeddings"]
H --> I["Claude Integration"]
end
%% External Databases Connected to Additional FastAPI
subgraph Databases["š¾ Connected Databases"]
K["Database 1"]
L["Database 2"]
M["Database 3"]
end
%% Connections Between Subgraphs
B --> D
F --> C
I --> C
J --> K
J --> L
J --> M
%% Output Components
B1["š No-Code AI Pipeline"]
B2["š¤ Multi-Agent System"]
B3["š” Context-Aware Responses"]
%% Linking Outputs
C --> B1
F --> B2
I --> B3
%% Class Assignments
class A,B,C,D,E,F,J process
class G,H,I,K,L,M data
class B1,B2,B3 output
An AI workflow pipeline with a drag-and-drop interface, enabling users to integrate data from existing databases, connect multiple LLMs with custom system prompts (agentic), and visualize the overall system. Developed with React and FastAPI, it provides a clean and intuitive frontend for building and understanding complex workflows.
flowchart TD
%% Define Styles
classDef input fill:#e3f2fd,stroke:#1565c0,stroke-width:2px
classDef process fill:#f3e5f5,stroke:#6a1b9a,stroke-width:2px
classDef storage fill:#e8f5e9,stroke:#2e7d32,stroke-width:2px
classDef output fill:#fff3e0,stroke:#e65100,stroke-width:2px
%% Document Processing Subgraph
subgraph PREP["š Document Processing"]
A[Upload Files] --> B[Hash Check]
B -->|New File| C[Custom Chunking]
B -->|Existing File| D[(Stored Embeddings)]
C --> E[Token Counting]
E --> F[Cost Estimation]
F --> G[Generate Embeddings]
G --> D
end
%% Query Processing Subgraph
subgraph QUERY["š Query Processing"]
H[User Question] --> I[Embed Query]
I --> J[Vector Search]
D --> J
J --> K[Fetch References]
K --> L[Generate Response]
L --> M[Citations & Sources]
end
%% Connections Between Subgraphs
G -.-> D
%% Class Assignments
class A,H input
class B,C,E,F,G,I,J,K,L process
class D storage
class M output
A Streamlit-based RAG chat interface that retrieves results with references using Titan v2 for embeddings and Claude for response generation.
Core Domains
Select a domain to explore related projects and portfolios:
