Skip to content
View mehakgoel20's full-sized avatar
3 followers · 5 following

Block or report mehakgoel20

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mehakgoel20/README.md 
╔══════════════════════════════════════════════════════════╗
║         MEHAK GOEL  //  AI/ML ENGINEER                   ║
║         Building production-grade GenAI systems          ║
╚══════════════════════════════════════════════════════════╝

LinkedIn Gmail Open to Internships Immediate Joining

Pre-final year @ Sharda University · AI/ML · Available immediately

$ whoami

mehak = {
    "role"       : "AI/ML Engineer | GenAI Builder",
    "university" : "Sharda University — B.Tech CSE (AI/ML), CGPA 8.0",
    "focus"      : ["RAG Systems", "LLM Engineering", "Fintech AI"],
    "status"     : "🟢 Actively seeking AI/ML internships — fintech & GenAI",
    "certified"  : [
        "Oracle OCI Generative AI Professional (2025)",
        "Oracle OCI AI Foundations (2025)"
    ]
}

I build production-grade AI systems — not just notebooks.
CredLens AI ships with audit logs, evaluation harnesses, and Docker.
FinQA RAG ships with refusal logic, Redis caching, and observable cost controls.

$ ls ./flagship-projects

Project Stack What it does
CredLens AI FastAPI · ChromaDB · RAG · Docker Explainable loan decisioning — policy-grounded RAG, deterministic underwriting rules, cryptographic hash-chained audit logs, 30-case eval harness.
FinQA Production RAG Qdrant · BM25 · Redis · FastAPI Hallucination-safe financial QA — hybrid retrieval, refusal logic, symbolic arithmetic execution, Redis caching, observable cost controls.
Insurance Prediction Pipeline Scikit-learn · Streamlit · Docker End-to-end ML pipeline predicting medical insurance charges — preprocessing, feature engineering, trained artifacts, deployed with Streamlit.
BBC News Classifier NLP · Scikit-learn · Classical ML News classifier on a self-engineered dataset — 42,000 raw RSS articles cleaned, deduplicated, and labelled from scratch.

$ cat ./tech-stack.json

{
  "languages"   : ["Python", "SQL", "JavaScript"],
  "ml_ai"       : ["LangChain", "RAG", "ChromaDB", "Qdrant", "FAISS",
                   "Scikit-learn", "Pandas", "NumPy"],
  "llm_infra"   : ["FastAPI", "Redis", "BM25", "Sentence Transformers",
                   "OpenAI API", "Anthropic API"],
  "devops"      : ["Docker", "Railway", "Git", "Linux"],
  "cloud"       : ["Oracle OCI", "GCP (learning)"]
}

$ cat ./what-i-build

[1]  RAG pipelines grounded in real constraints — no hallucinations, no magic
[2]  Credit decisioning with explainability, audit trails & policy alignment
[3]  ML systems: raw messy data → production API
[4]  GenAI tools that are observable, testable, and deployable

Visitor Count

Open to AI/ML internships across any domain — reach out at gmehak350@gmail.com

Pinned Loading

  1. credlens-ai credlens-ai Public

    Explainable AI loan decision system using RAG, ML risk scoring, policy retrieval, evidence verification, and real-time monitoring dashboard built with FastAPI, ChromaDB, and Docker.

    Python 1

  2. financial-rag-system financial-rag-system Public

    Production-grade Retrieval-Augmented Generation (RAG) system built on financial documents (FinQA dataset). The system is designed to be hallucination-safe, cost-efficient, and observable.

    Python

  3. Insurance-prediction-pipeline Insurance-prediction-pipeline Public

    End-to-end ML pipeline for predicting medical insurance charges using Python, Streamlit, and Docker. Includes complete preprocessing, scaling, engineered features, trained model artifacts, and a fu…

    Jupyter Notebook

  4. bbc-news-classification-pipeline bbc-news-classification-pipeline Public

    This project builds a news category classifier using classical NLP + ML techniques on a custom-engineered BBC News dataset. Starting with 42,000 raw RSS articles, I cleaned, deduplicated, and label…

    Jupyter Notebook