Skip to content
YG
Portfolio · Vol. 01Inbox open

YashGupta, AI/ML Engineer, Data Analyst, Oracle Certified Professional

AI/ML Engineer · Data Analyst · Oracle Certified Professional

I turn complex data into clear decisions, and build AI that automates the busywork.

85–90%

of the bed-allocation and patient-tracking workflow automated in a hospital management system I built. See the case study

View workGet in touch
§01Selected work

Selected work

Case studies in forecasting, computer vision, automation and language, each written up as problem, approach and result.

04 case studies

No. 01Case study

Sales Prediction & ROI Optimizer

Fig. 1.1 Sales forecast with an uncertainty band, and the features driving itIllustrative
Problem
Businesses need to know where a marketing budget will pay off before they commit it.
Approach
Real-time sales forecasting with scikit-learn, dynamic ROI calculation, and feature-importance visualization that shows what drives each forecast, delivered as a Streamlit app.
Result
Budget allocation becomes a measurable decision: live forecasts, dynamic ROI and the drivers behind every prediction, in one platform.
  • Python
  • Scikit-learn
  • NLPaas
  • Streamlit
No. 02Case study

Emotion Recognition System

Fig. 1.2 Affect scores across seven emotions, current frame against the previous oneIllustrative
Problem
Scoring how a user feels in real time means making sense of high-dimensional input the moment it arrives.
Approach
A multimodal system built with OpenCV and TensorFlow, using self-learning algorithms to process high-dimensional data on the fly.
Result
Real-time affect scores, surfaced through an intuitive emoji-based interface anyone can read at a glance.
  • Python
  • OpenCV
  • TensorFlow
  • Streamlit
No. 03Case study

Automated Hospital Management

Fig. 1.3 Share of the workflow automated, one cell per percentage pointIllustrative
85–90%workflow automated
Problem
Bed allocation and patient tracking were run as manual workflows.
Approach
A Python and MySQL system that automates allocation and tracking, with AI chatbots built on the OpenAI API handling system automation and reporting.
Result
85–90% of the bed-allocation and patient-tracking workflow automated.
  • Python
  • MySQL
  • OpenAI API
No. 04Case study

Email Spam Detector

Fig. 1.4 Spam and ham separated by a learned decision boundaryIllustrative
Problem
Separating spam from legitimate email, and knowing how confident the model is about every call.
Approach
An NLP pipeline with scikit-learn classifiers that labels each email as spam or ham, scores its confidence and flags anomalies.
Result
A Streamlit app that classifies email with a confidence score and keeps a session history of every check.
  • Python
  • NLP
  • Scikit-learn
  • Streamlit
§02Experience

Five internships, eighteen months.

From data analytics through machine learning to generative AI, at IBM CSRBOX, IBM SkillsBuild and Oasis Infobyte.

  • Data Analytics Intern & Team Lead · IBM CSRBOX
  • Machine Learning Intern · IBM SkillsBuild (CSRBOX)
  • Prompt Engineering Intern · IBM CSRBOX
  • AI & Machine Learning Intern · Oasis Infobyte
  • Data Science Intern · Oasis Infobyte
Fig. 2 Internships on one axis, Jun 2024 to Nov 2025. Select a bar to jump to the role.

  1. Oct 2025 – Nov 2025

    02 months · Remote

    Oasis Infobyte

    Data Science Intern

    • Demonstrated exceptional organization and outstanding project delivery
    • Developed “QuantumAd Prophecy,” an AI platform for effectively prioritizing advertising budgets
    • Engineered a robust Email Spam Detection system using NLP techniques
    • Conducted comprehensive unemployment analysis in India using Plotly
    • Machine Learning
    • NLP
    • Plotly
    • Data Visualization

  2. Jan 2025 – Feb 2025

    02 months · Remote

    Oasis Infobyte

    AI & Machine Learning Intern

    • Engineered a real-time Feedforward Neural Network achieving 89% initial accuracy
    • Implemented self-learning algorithms that ingest production data to continuously retrain models
    • Optimized Generative AI workflows on IBM Watson and ChatAPI, boosting content generation speed by 48%
    • Structured domain-specific LLM tuning and reduced inference response time
    • Machine Learning
    • Generative AI
    • IBM Watson

  3. Dec 2024 – Jan 2025

    02 months · Remote

    IBM CSRBOX

    Prompt Engineering Intern

    • Optimized Generative AI workflows on IBM Watson and ChatAPI
    • Increased content generation speed by 48%
    • Structured domain-specific LLM tuning
    • Managed the end-to-end pipeline: data handling, SQL querying, and visualization
    • Prompt Engineering
    • Generative AI
    • IBM Watson
    • SQL
    • Data Visualization

  4. Oct 2024 – Nov 2024

    02 months · Remote

    IBM SkillsBuild (CSRBOX)

    Machine Learning Intern

    • Worked on the “Mastering Data with Machine Learning” program
    • Deployed predictive models that enhanced operational forecasting efficiency
    • Led a team to analyze sustainability datasets
    • Delivered insights that improved operational forecasting by 20%
    • Machine Learning

  5. Jun 2024 – Aug 2024

    03 months · Remote

    IBM CSRBOX

    Data Analytics Intern & Team Lead

    • Led a team to analyze sustainability datasets
    • Delivered insights that improved operational forecasting by 20%
    • Managed the end-to-end pipeline: data cleaning, SQL querying, and visualization
    • SQL
    • Data Visualization
§03Skills

The toolkit

Grouped by discipline and cross-referenced against the work: every mark is a project, a role or a credential where the skill shows up.

30 tools · 60 cross-references

ProjectRoleCredentialSelect a skill to see where it shows up

Machine learning & AI

10

Also in the kitPyTorch, Keras

Data & analytics

08

Also in the kitPandas, NumPy, Tableau, Power BI, PostgreSQL

Tools & cloud

09

Also in the kitAWS, Docker, Git & GitHub, FastAPI, React

Languages

03

Also in the kitJavaScript

§04Certifications

On the record

Three Oracle Professional certifications lead a register of sixteen, from Oracle University, IBM SkillsBuild, NPTEL, Forage and more.

16 credentials

  • Oracle UniversityProfessional

    OCI Generative AI Professional

    LLMs · Prompt Engineering · OCI GenAI Service · Fine-tuning

    2025
  • Oracle UniversityProfessional

    OCI Data Science Professional

    Machine Learning · Model Deployment · OCI Data Science · AutoML

    2025
  • Oracle UniversityProfessional

    AI Vector Search Professional

    Vector Database · RAG · Semantic Search · Embeddings

    2025
Fig. 4 Credentials by issuer. Select one to filter the register.

Showing all 16 credentials.

§05ContactInbox open

Got messy data or a manual process? Let's talk.

Whether you have a question, a project proposal, or just want to connect, my inbox is always open.

Write directly

yashgupta.fdr@gmail.com

Or send a message