Available for opportunities

Hi, I'm

Aman Kumar

Building production AI systems, LLM-powered applications, and adaptive ML pipelines. Currently at Sandisk India — open to challenging roles across AI, data, and software engineering.

View My Work Download Resume
About

Who I Am

Aman Kumar

Machine Learning Engineer with hands-on experience building adaptive AI pipelines, LLM-based agent systems, and production MLOps infrastructure.

I enjoy working across the stack — from designing and training models to deploying and monitoring them in production. Comfortable diving deep into research problems or shipping robust, scalable systems. I build things that work and can be explained clearly.

Bengaluru, Karnataka, India
amankumar24052001@gmail.com
MTech Integrated Data Science — VIT Vellore · CGPA 8.1
Download Resume
Career

Experience

Sandisk India

Machine Learning Engineer

Sept 2025 — Present Current
  • Enhanced multi-corner timing predictions for the SWIFTECO framework: MAE 8 ps (0.008), 20% over-estimation, under-estimation ≤0.005 units
  • Reduced timing analysis turnaround from over a day to 47 minutes, accelerating ECO validation cycles and throughput
  • Processed large PrimeTime reports (10+ GB/rpt) with scalable pipelines for parsing, feature extraction, and automated validation
  • Designed Stream IP and RTL Lint Agents enabling natural-language queries and automated RTL cleanup, reducing manual intervention

Hyperbots

Applied ML Engineer

Jun 2025 — Sept 2025
  • Productionized MLOps pipelines across AWS/GCP with Docker & Jenkins, ensuring reliable deployment of ML and LLM-based agent systems
  • Built intelligent document processing pipeline (Gradio + OCR + HIL) for RAG-style retrieval and annotation workflows, reducing manual validation
  • Implemented Slack-integrated ML feedback loop improving invoice conflict detection and enabling real-time adaptive predictions
  • Developed real-time monitoring dashboards in React for pipeline health, model performance, and workflow accountability

Western Digital (Sandisk India)

AI-ML Intern

Sep 2024 — Jun 2025
  • 55% improvement in STA timing consistency using optimization-driven ML models, reducing design iteration cycles
  • Automated IR drop and PrimeTime report analysis in Python, strengthening verification reliability
Work

Projects

Featured Project

Adaptive Macro Regime Detection & ETF Rotation

3-state Gaussian HMM trained on 12 FRED macro series — yield curve, credit spreads, CPI, unemployment — to detect Expansion / Stagnation / Contraction regimes and dynamically rotate a 6-asset ETF portfolio (SPY / TLT / GLD / LQD / HYG / BIL). 241-month walk-forward OOS backtest with zero look-ahead bias.

0.937 Sharpe (SPY: 0.760)
+3.4% 2008 GFC (SPY: −46%)
−21% Max DD (SPY: −50.8%)
Python hmmlearn FRED API Walk-Forward HMM GARCH Quantitative Finance
Live Dashboard View Code
Talk to DB

Talk to DB

Natural language querying over relational databases using LLM + LangChain pipelines. Schema-aware prompting with validation and adaptive feedback loops.

Streamlit LangChain RAG SQL
Code
Market Analysis Wizard

Market Analysis Wizard

AI-powered market research using Gemini and SerpAPI to generate comprehensive reports with dynamic visualizations and SWOT analysis.

Gemini AI SerpAPI Python
Code
COVID-19 Dashboard

COVID-19 Automated Dashboard

Real-time data pipeline converting COVID-19 datasets to KPIs and interactive Tableau visualizations for executive reporting.

Tableau Data Pipeline KPIs
Live
Fake News Detection

Fake News Detection

Robust detection system using ML ensembles — Logistic Regression, Decision Trees, Gradient Boosting, and Random Forest — with comparative evaluation.

scikit-learn NLP Ensemble ML
Report
Know Your Website

Know Your Website

Web threat detection application — classifies web requests using ML models trained on network traffic features to flag malicious patterns.

ML Security Flask
Code
Snake Master

Snake Master

Advanced Snake game with modern mechanics, progressive difficulty, and responsive design. Built with vanilla HTML5 Canvas — no frameworks.

HTML5 Canvas JavaScript CSS3
Play
Expertise

Technical Skills

Languages

Python SQL Java Tcl

ML / AI Frameworks

TensorFlow Keras scikit-learn PyTorch Hugging Face LangChain

Generative & Adaptive AI

LLMs RAG NLP ChromaDB Pinecone FinBERT

MLOps & Cloud

Docker Kubernetes Jenkins AWS GCP GitHub Actions

Backend & APIs

FastAPI Flask React.js Streamlit Gradio

Data & Analytics

Pandas NumPy Plotly Matplotlib A/B Testing EDA

Tools & Platforms

Git Linux PostgreSQL Power BI Jupyter
Background

Education

Vellore Institute of Technology, Vellore

5-Year Integrated

MTech Integrated Data Science

Sep 2020 – Jul 2025  ·  CGPA: 8.1 / 10

Machine Learning Deep Learning NLP Statistical Modelling Big Data Data Structures

Army Public School, Ranchi

CBSE XII

Completed 2020  ·  93.8%

Certifications

Achievements

Artificial Intelligence Analyst

IBM

Introduction to Generative AI

Google

Data Science Virtual Experience

British Airways · Forage

Statistics (Basics, Probability & Datasets)

LinkedIn Learning
Let's Talk

Get In Touch

Open to discussing ML engineering roles, quant research positions, or interesting collaborations. Drop me a message — I read every email.

amankumar24052001@gmail.com
+91 9065939752 Bengaluru, Karnataka, India
Document

Resume