⚡ GEDENDHAR SIVAKUMAR

AI/ML Engineer • Python Developer • AWS Certified

---

🧠 Profile Overview

Results-driven AI/ML Engineer and Python Developer specializing in enterprise-grade Generative AI systems, Multi-Agent workflows, and robust NLP/RAG architectures. Experienced in deploying containerized solutions on AWS, Render, Vercel, and Streamlit Cloud infrastructures, building high-throughput document intelligence pipelines, and securing LLM applications through adversarial red-teaming.

💻 Core Tech Stack

🤖 Generative AI & Orchestration

🧠 AI/ML & Natural Language Processing

📂 Vector Search & Document Intelligence

☁️ Deployment & Cloud Hosting

🚀 AI Development Partners & Large Language Models

Autonomous governance multi-agent swarm that audits LLMs against global regulatory standards.

• Multi-Agent Swarm: Parallelized evaluations via LangGraph to audit Legal, Privacy, Bias, and Red-Teaming vulnerabilities simultaneously.
• Adversarial Probing: Simulates prompt injections and jailbreaks to measure target model resilience.
• Export Pipeline: Renders scores and compliance parameters into an official AI Compliance Passport (PDF).
• 🛠️ Tech Stack: Python, LangGraph, Groq Cloud (Llama 3.3), FAISS, Streamlit, Docker, FPDF2.
• 🔗 View Project Repository

graph LR
    A["⚙️ Ingest Application Config"] --> B["🧭 LangGraph Orchestration"]
    B --> C1["🏛️ Legal Agent (FAISS RAG)"]
    B --> C2["🔒 Privacy Agent (PII Scan)"]
    B --> C3["⚖️ Bias Agent (Toxicity Eval)"]
    B --> C4["🔴 Red-Team Agent (Jailbreak)"]
    C1 & C2 & C3 & C4 --> D["👔 Chief Compliance Officer"]
    D --> E["📄 AI Passport PDF"]
    style B fill:#1c3c3c,stroke:#333,stroke-width:2px,color:#fff
    style D fill:#4a154b,stroke:#333,stroke-width:2px,color:#fff

Automated data analytics suite that cleans data, generates dashboards, extracts insights, and computes forecasts in a single click.

• One-Click Analytics: Raw file upload automatically starts cleaning, structural standardizations, and error rectifications.
• Predictive Modeling: Ingests historical data, fits time-series trends, and plots forecasts via interactive Plotly widgets.
• 🛠️ Tech Stack: Python, Pandas, Streamlit, Scikit-Learn, Plotly.
• 🔗 View Project Repository

graph LR
    In["Raw Data (CSV/Excel)"] --> Clean["`🧹 Auto-Data Cleaning (Pandas)`"]
    Clean --> Forecast["`📈 Predictive ML Forecasting (Scikit-Learn)`"]
    Forecast --> Insights["`💡 Insights Generator (GenAI)`"]
    Insights --> UI["📊 Live Plotly Dashboard"]
    style Clean fill:#121013,stroke:#333,color:#fff
    style UI fill:#ff4b4b,stroke:#333,color:#fff

Secure document redaction tool designed to protect whistleblowers and sensitive sources.

• Source Safety: Scans images or raw text PDFs using OCR, runs NLP Named Entity Recognition to locate identity markers, and flattens outputs.
• Security Matrix: Rates leaking risks and formats high-fidelity black-bar redactions.
• 🛠️ Tech Stack: Python, Transformers, NLP NER, Streamlit, Document Parsing.
• 🔗 View Project Repository

graph LR
    Doc["Whistleblower Doc"] --> NER["🔍 Named Entity Recognition (NER)"]
    NER --> Mask["🔒 Token Masking & Redaction"]
    Mask --> Risk["📉 Security Safety Score Assessment"]
    Risk --> Out["📄 Redacted Safe PDF"]
    style NER fill:#8a2be2,stroke:#333,color:#fff
    style Risk fill:#e94560,stroke:#333,color:#fff

Custom Claude capabilities implementing advanced prompt-engineering frameworks for structured generation.

• Prompt Engineering Standard: Standardizes system message architectures to extract predictable JSON variables.
• 🛠️ Tech Stack: Claude API, System Prompts, Prompt Engineering, JSON Schema.
• 🔗 View Project Repository

graph LR
    Raw["User Request"] --> RCTFCE["📐 Apply RCTFCE Engine"]
    RCTFCE --> System["📝 Compile System Message"]
    System --> Inference["🤖 Anthropic Claude Execution"]
    Inference --> Format["🎯 Strict Structured JSON Output"]
    style RCTFCE fill:#46e3b7,stroke:#333,color:#000
    style Format fill:#0467df,stroke:#333,color:#fff

Autonomous AI observability platform for real-time root cause analysis and infrastructure self-healing.

• Self-Healing loops: Observes cluster pods, reads log lines, runs RCA, and executes recovery scripts autonomously.
• 🛠️ Tech Stack: Python, LangChain, Kubernetes API, Groq, Logging Observability.
• 🔗 View Project Repository

graph LR
    Metric["📡 Cluster Telemetry Logs"] --> Detect["🔍 Anomaly Detection Engine"]
    Detect --> Reason["🧠 LLM Root Cause Analysis (RCA)"]
    Reason --> Recover["🛠️ Remediation Orchestrator"]
    Recover --> Repair["✅ Self-Healing Actions executed"]
    style Reason fill:#1C3C3C,stroke:#333,color:#fff
    style Recover fill:#ff5722,stroke:#333,color:#fff

Generative AI log analysis conversational agent for developers and systems engineers.

• Stacktrace Ingestion: Decodes multi-line runtime crashes, queries model paths, and outputs exact line-number modifications.
• 🛠️ Tech Stack: FastAPI, Groq API (Llama 3), LangChain, Python.
• 🔗 View Project Repository

graph LR
    Logs["App Logs Ingest"] --> Fast["⚡ FastAPI Endpoint"]
    Fast --> LC["⚙️ LangChain Extraction Node"]
    LC --> Model["🤖 Llama 3 (Groq API) Analysis"]
    Model --> Suggest["💡 Root Cause + Fix Script"]
    style Fast fill:#00d2ff,stroke:#333,color:#000
    style Model fill:#8a2be2,stroke:#333,color:#fff

Predictive analytics engine for Formula 1 Grand Prix classification and outcome forecasting.

• Race Predictive Flow: Maps qualifying telemetry, weather parameters, and driver track metrics to model race grids.
• 🛠️ Tech Stack: Python, Pandas, Scikit-Learn, LightGBM, Data Engineering.
• 🔗 View Project Repository

Generative AI tool for resume optimization, automated matching, and career positioning.

• Resume Alignment: Computes cosine similarities between candidate files and job specs, outputting structured improvement plans.
• 🛠️ Tech Stack: Python, OpenAI ChatGPT, LangChain, RAG.
• 🔗 View Project Repository

graph LR
    Resume["📄 CV (PDF/DOCX)"] & JD["💼 Job Description"] --> Parse["⚙️ Document Parser"]
    Parse --> Embedding["🔢 Dense Vector Embeddings"]
    Embedding --> Cosine["📏 Cosine Similarity Matching"]
    Cosine --> LLM["🤖 LLM Refiner (RAG Mode)"]
    LLM --> Out["⚡ Optimised Resume + Interview Tips"]
    style Parse fill:#0467df,stroke:#333,color:#fff
    style LLM fill:#10a37f,stroke:#333,color:#fff

• Scalable NLP Pipelines: Architected high-performance natural language processing pipelines in Python to process unstructured text at scale.
• Robust Backends: Developed, tested, and optimized microservices and RESTful API endpoints for core enterprise products.
• AWS Infrastructure: Deployed and managed serverless and containerized applications using Amazon Web Services (AWS) ECS, Lambda, and S3.
• Workflow Automation: Automated internal developer loops and CI/CD procedures, boosting engineering efficiency by 30%.

• 🎓 AWS Certified Solutions Architect – Associate (Amazon Web Services)
• 📜 Python Programming (Google | Coursera)
• 💼 Agile Methodology Virtual Experience (JPMorgan Chase)
• 🤖 Claude in Action (Anthropic Claude Frameworks)