Diwas Parashar // Portfolio

Engineering the
Next Generation of AI.

Pursuing

M.Sc. Artificial Intelligence & Data Analytics student at Amity University, Noida. Focused on developing end-to-end Machine Learning pipelines, deep sequence classifiers, and predictive time-series models.

diwas_pipeline_monitor.sh
TRAINING MODEL [DIWAS-GPT-v1]

> epoch 15/50: loss=0.104 acc=91.2%

> epoch 30/50: loss=0.038 acc=96.8%

> epoch 45/50: loss=0.015 acc=98.4%

> process finished with code 0 (success)

Global F1-Score: 98.42%

About Me

> cat diwas_profile.json
{
  "name": "Diwas Parashar",
  "location": "Noida, India",
  "education": "M.Sc. AI & Data Analytics",
  "university": "Amity University, Noida",
  "undergrad": "BCA, Bharati Vidyapeeth",
  "focus": ["Machine Learning", "Deep Learning", "Time-Series ML"],
  "status": "Seeking summer internships & full-time roles"
}
5+
Core ML Projects
89%
Max ROC-AUC

Bridging Algorithmic Logic & Data Insights

I am an M.Sc. student specializing in Artificial Intelligence and Data Analytics at Amity University. With a robust undergraduate foundation in Computer Applications from Bharati Vidyapeeth University, Pune, I specialize in constructing predictive machine learning frameworks, tuning sequence classification networks, and analyzing structured business data.

My engineering approach centers on modular programming design in Python, utilizing mathematical concepts for regression/classification, feature extraction, and performance evaluation. I am highly motivated to implement interpretable AI models that automate operational intelligence and support predictive decision systems.

  • ML Pipelines & Evaluation
  • Time-Series Forecasting
  • Deep Sequence Modeling
  • Explainable AI (SHAP)

Technical Expertise

Machine & Deep Learning

Python Programming 95%
Feature Engineering & Evaluation 90%
ML Classification & Regression 88%
Deep Learning (LSTMs, TensorFlow) 82%

Data & Analytics

Pandas, NumPy & Wrangling 92%
SQL Database & Schema Design 85%
Exploratory Data Analysis (EDA) 90%
Data Visualizations & Reports 85%

Tools & Core Concepts

Git & Version Control 85%
APIs & Service Connections 78%
Firebase Backend Integration 75%
Cloud Fundamentals & LaTeX 70%

Technical Projects

Python Scikit-learn Supervised ML

Parkinson's Disease Severity Analysis

Designed an end-to-end ML pipeline using 200+ biomedical voice samples to predict Parkinson's disease presence and severity. Engineered key features and compared multiple models (Logistic Regression, Random Forest, SVM) via cross-validation, successfully optimizing prediction accuracy from 78% to 86%.

XGBoost SHAP SMOTE

Customer Churn Prediction with Explainability

Built an interpretable churn prediction model on a dataset of 7,000+ customers. Addressed class imbalance using SMOTE and achieved an ROC-AUC score of 0.89. Utilized SHAP values to identify key churn drivers and optimized decision thresholds to balance precision and recall.

Time-Series ML Anomaly Detection Rolling Windows

AI-Based Predictive Maintenance System

Developed a time-series predictive model using simulated sensor data (10,000+ records) to forecast smart manufacturing equipment failures. Extracted rolling window features and trend lines, achieving 84% anomaly prediction accuracy and reducing downtime by 25%.

TensorFlow PyTorch LSTM Networks

Deep Learning Text Classification

Constructed an LSTM-based deep classifier trained on 20,000+ text samples to capture long-term sequential dependencies. Implemented custom tokenization, padding, and embedding layer architectures, achieving an 88% classification accuracy (a 10% improvement over baselines).

NLP Research Sarcasm Detection Literature Review

Decade Review of Sarcasm Detection (2015-2025)

Conducted a systematic literature review analyzing 50+ papers spanning traditional ML, deep learning, and transformer architectures. Developed a detailed taxonomy based on contextual feature engineering and identified critical research gaps in multilingual and cross-domain sarcasm detection.

Academic Journey

August 2025 - Present

M.Sc. in Artificial Intelligence & Data Analytics

Amity University, Noida

Advanced studies in neural optimization, machine learning modeling, structured data mining, predictive statistics, and big data configurations. Focusing on practical applications of deep learning pipelines.

October 2021 - June 2024

Bachelor's in Computer Applications (BCA)

Bharati Vidyapeeth University, Pune

Graduated with a CGPA of 6.9. Constructed strong core foundations in computer applications, programming paradigms (Java, Python), database management systems (SQL), and algorithm designs.

Co-Curricular Honors & Leadership

Society Operations & Strategic Directives

Bharati Vidyapeeth & TAFS MUN

- **President**, Music Society BVP Pune: Led performer divisions and directed major events.
- **Secretary General / Co-Secretary General**, TAFS Model United Nations (2019-2020): Led event coordination, directed logistics, and handled delegate operations.
- **NCC Air Wing 'C' Certificate** Holder.

Interact With DiwasAI Chatbot

To demonstrate some AI design principles in a fun way, I have built a simulated neural assistant. Recruiters can click any of the programmed terminal queries on the right to receive immediate answers regarding my training, projects, or goals.

Responses are streamed directly in standard monospace console formatting.

diwas_assistant_v1.0.sh
ONLINE
> INITIALIZING DIWAS.AI INTERACTIVE CONSOLE...
> HELLO! I AM DIWAS'S AGENT. CLICK AN OPTION BELOW TO RETRIEVE PROFESSIONAL DATA STREAMS.

Let's Connect

I am open to discussions regarding summer internship positions, junior data analyst opportunities, or AI/ML collaborations. Reach out via email, social links, or submit a message through the dashboard.

Email Address

diwas.parashar@gmail.com

Telephone Connection

+91 9654039726

Primary Location

Noida, Uttar Pradesh, India