Aditya KumarSah
I build systems that hold up under real complexity — an analytics platform covering all 92 Nifty 100 companies, an AI agent with persistent memory, and the CS fundamentals behind both.
Engine
Comparison
A bit about how I work
Final-year B.Tech Computer Science student (Data Science specialization), graduating May 2027. Just wrapped up a Data Analyst internship at Bluestock Fintech, where I solo-architected a production-grade analytics platform across all 92 Nifty 100 companies.
I like problems with real edge cases — the kind where "handle outliers" turns into deciding what a debt-free company's D/E ratio should even mean. That instinct for structured, sprint-based execution — I shipped that project on a 6-sprint, 45-day roadmap — is what I bring to software engineering, not just analytics.
Coding Profiles
daily DSA practice, in JavaWhere I've worked
- Independently architected the N100 Financial Intelligence Platform — full-stack analytics covering all 92 Nifty 100 companies: ETL pipeline, financial ratio engine, stock screener, peer comparison module, dashboard, automated PDF reporting, and a FastAPI backend.
- Designed a composite data-quality scoring model with winsorization for outlier handling, plus custom logic for real edge cases — debt-free company D/E ratios, ICR passthrough, CAGR calculation.
- Built the end-to-end ETL pipeline ingesting, cleaning, and transforming financial data for 92 companies into analytics-ready outputs.
- Shipped a companion Mutual Fund ETL & Analytics Platform: pipeline, interactive Power BI dashboards, a 16-slide deck, and a written report.
- Built and shipped features across front-end and back-end of a live production PHP application.
- Worked inside a real engineering workflow — code review, bug triage, deployment — alongside the core dev team.
Selected work
Ordered for a systems-and-algorithms read first, then data engineering, then AI.
DepScan
In DevelopmentPython Dependency Analyzer CLI — maps a project's import graph, flags circular dependencies, and surfaces stale packages, straight from the terminal.
A direct build-out of my GSoC 2026 proposal to the Python Software Foundation. Parsing, graph algorithms, and CLI/DX design — the CS fundamentals behind developer tooling.
- AST-based import parser (Python
astmodule) - Dependency graph construction + cycle detection (DFS / topological sort)
- CLI with tree and JSON output modes
- Stretch: cross-reference PyPI for outdated/vulnerable packages
- Tests + CI, published to PyPI
N100 Platform
ShippedA full analytics platform covering every company in the Nifty 100 — architected and shipped solo across a 6-sprint roadmap during my Bluestock Fintech internship.
- Composite quality-scoring model with winsorization
- Custom edge-case logic: debt-free D/E, ICR passthrough, CAGR
Mutual Fund ETL & Analytics
ShippedEnd-to-end pipeline turning raw mutual fund data into an interactive Power BI story.
- Full extract-transform-load pipeline
- 16-slide stakeholder presentation + written report
AudienceIQ
ShippedAn AI analytics agent that actually remembers you.
- Custom "Hindsight" memory architecture
- Persistent context recall across sessions
Tools & fundamentals
Languages
CS Fundamentals / DSA
Backend & Systems
Data Science & ML
Data Engineering
Visualization
Cloud & Tools
Education & certifications
Let's build
something worth shipping.
Open to Software Engineer and Data Analyst conversations. I reply fast.