Projects | Veomn

My Projects

RoveMiles — Data Analysis + Financial Modeling Work

Goal:

Built end-to-end predictive analytics and financial modeling system for RoveMiles.

Problem Statement:

Analyzed large transactional datasets, engineered financial metrics, and contributed to dashboards for revenue modeling and key business KPIs.

Result:

Delivered a comprehensive financial model, identified high-impact churn and profitability drivers, and built dashboards for revenue and unit-economics analysis.

Approach / Methodology:

Built a multi-sheet financial forecasting model covering GMV, take rate, redemption cost, partner payouts, revenue curves, and profit margins.
Performed scenario and sensitivity analysis on mile issuance/redemption behavior to evaluate profitability.
Analyzed partner, redemption, and customer behavior datasets to identify churn factors and optimize pricing.
Created dashboards and tools for KPI tracking and revenue modeling.

Key Learnings:

Business-focused machine learning
Working with ambiguous real-world datasets
Integrating financial models with predictive analytics

Skills:

Excel financial modeling (GMV, take rate, unit economics), data analysis, scenario forecasting, partner-pricing modeling, revenue/cost analysis, dashboard creation

Real-Time Customer Behavior Analytics Pipeline (AWS)

Goal:

Turning live data into instant business insight.

Problem Statement:

Businesses need immediate visibility into customer interactions, but traditional batch-processing systems cause delays in detecting anomalies, trends, and engagement patterns.

Results / Visuals:

Cut event-to-dashboard latency to under 2 minutes
Fixed Redshift cost spike (~$1.3k) and optimized compute use
Enabled BI team to visualize customer trends in near-real-time
Fully fault-tolerant streaming architecture

Approach / Methodology:

Designed full architecture: Kinesis → Lambda → S3 → Glue Crawler → Glue ETL → Glue DQ → Redshift → QuickSight
Built PySpark ETL scripts for cleaning, partitioning, transforming event data
Set up Glue Data Catalog tables, schema inference, and automated crawlers
Configured Redshift clusters, IAM roles, networking, VPC, subnets
Implemented AWS Budgets + cost controls after diagnosing Redshift overages
Created QuickSight dashboards for funnels, anomalies, activity timelines

Key Learnings:

Deep practical understanding of cloud data engineering architecture
Schema evolution, Glue DQ, IAM policy debugging, Redshift performance tuning
How real-time ETL differs from batch pipelines

Skills:

AWS Kinesis, AWS Lambda, AWS Glue, Glue DQ, Redshift, S3, Athena, CloudWatch, PySpark, VPC, IAM

GitHub I View Report

AI-Powered Retrieval-Augmented Generation System

Goal:

Offline AI for secure, collaborative insight.

Problem Statement:

Teams handling sensitive or confidential data need a locally hosted LLM solution that runs fully offline while providing accurate retrieval, high-quality responses, and support for multi-agent collaboration.

Results / Visuals:

● Reduced API cost by 100% via full offline inference

● Achieved faster hybrid retrieval through graph + vector search

● Delivered a modern UI suitable for internal research workflows

Approach / Methodology:

Built a full RAG architecture using Microsoft GraphRAG for structured retrieval
Integrated AutoGen agents for multi-agent reasoning
Used Ollama to run local models (Mistral, Nomic embeddings)
Added Lite-LLM proxy to enable OpenAI-style function calling with local models
Designed an interactive Chainlit UI supporting multi-threaded chats
Generated embeddings and stored them in local vector stores

Key Learnings:

Differences between graph-based and dense retrieval
Offline inference tradeoffs, local model memory/latency handling
Multi-agent orchestration patterns

Skills:

Python, Chainlit, Microsoft GraphRAG, AutoGen, Lite-LLM, Ollama, Mistral, Nomic Embeddings, FAISS

GitHub

ChurnGuard — Customer Churn Prediction ML Platform

Goal:

ML-based churn prediction with +15% accuracy improvement.

Problem Statement:

Businesses need tools to forecast customer churn and proactively address retention weaknesses

Results / Visuals:

● 15% accuracy increase after tuning

● Clear ranking of features affecting churn risk

● Delivered a dashboard-ready dataset of churn probabilities

Approach / Methodology:

Cleaned and normalized customer datasets
Feature engineering + outlier detection
Used Logistic Regression, Random Forests, Neural Networks, Gradient Boosting
Tuned hyperparameters to maximize predictive score
Evaluated with accuracy, ROC-AUC, precision-recall

Key Learnings:

ML model experimentation workflows
Importance of preprocessing and data quality
Tradeoffs between different classification models

Skills:

Python, Pandas, Numpy, Scikit-Learn, Matplotlib, Jupyter

GitHub

Gene Expression Visualizer- RNA-seq & Microarray Analysis

Goal:
Simplify gene-expression visualization for high-dimensional datasets.

Problem Statement:
RNA-seq/microarray data contain thousands of genes, requiring clustering, dimensionality reduction, QC checks, and publication-quality plots — tasks that otherwise need extensive coding.

Approach / Methodology:

Loaded & normalized expression matrices
Performed PCA & hierarchical clustering
Generated correlation and QC plots
Automated report creation via unified API

Results / Visuals:

Clustered heatmaps
PCA sample-relationship plots
Correlation matrices
Expression distribution box plots* One-command automated reporting

Key Learnings:

Handling high-dimensional biological data
Normalization & QC strategies
Accessible visualization design

Skills:

Python, Pandas, NumPy, Matplotlib, Seaborn, Scikit-Learn

GitHub

DNA Analysis Toolkit — Comprehensive Sequence Analysis

Goal:

Provide an all-in-one DNA analysis tool for students and researchers.

Problem Statement:

DNA analysis typically requires multiple tools for GC content, ORFs, codon usage, translation, restriction sites, and molecular properties — making workflows slow and fragmented.

Approach / Methodology:

FASTA parsing and standardized structure
GC, ORF, codon usage, translation pipelines
Restriction-site scanning
Visualization with Matplotlib/Seaborn
Memory-efficient batch workflows

Results / Visuals:

GC-content plots, ORF maps
Codon-usage heatmaps
Restriction-site visualizations
Molecular weight/Tm calculations
Batch processing for 1000+ FASTA sequences

Key Learnings:

DNA analysis fundamentals
BioPython workflows
Scientific visualization best practices

Skills:

Python, Biopython, Pandas, Matplotlib, Seaborn

GitHub

BLAST Result Parser — Automated Sequence Alignment Analysis

Goal:
Automate BLAST alignment parsing and filtering for faster biological insight.

Problem Statement:

BLAST outputs are large and difficult to interpret manually. Researchers need simpler tools to filter hits by e-value, identity, and bit score, extract top matches, and export results for downstream analysis.

Approach / Methodology:

Parsed XML/tabular BLAST with Biopython
Standardized alignment fields
Built chainable filtering pipeline
Computed statistics via pandas
Provided multi-format export

Results / Visuals:

Clean filtering by e-value, identity %, bit score
Summary statistics for alignment quality
Query-grouped and top-hit reports
CSV/JSON exports for further analysis

Key Learnings:

BLAST metric interpretation
Scientific data-pipeline design
Usable API design for researchers

Skills:

Python, Biopython, Pandas, JSON/CSV

GitHub