Open to opportunitiesAhmedabad, IndiaData & Digital Systems Analyst — Elixir FoundationB.Tech IT · Silver Oak
Turningnoise intosignal.
I'm Ruchir — a data analyst who takes messy, half-structured, real-world data and leaves it as something a team can actually decide with. Dashboards that survive real use, pipelines that don't break at 2am, and full-stack tools built with equal parts curiosity and discipline.
What pulled me into IT wasn't the syllabus — it was wanting to know what's actually happening behind the screen. A spreadsheet full of numbers is just noise until you find the story it's telling. Then it's a decision someone can act on.
I'm based in Ahmedabad, pursuing my B.Tech in Information Technology at Aditya Silver Oak Institute of Technology after finishing my Diploma in IT in 2026. Since June 2026 I've been the Data & Digital Systems Analyst at Elixir Foundation, an NGO where the datasets aren't practice sets — they're beneficiary and program records that decide where real support goes.
Outside that, I build: a client-side RAG engine, a cybersecurity IOC dashboard, government-data analyses I publish honestly (including the ones where my hypothesis died), and utilities I actually use. I'd rather hear hard feedback early than ship something half-right.
How I work
Growth mindset
Learning something new is the default state, not a phase.
Problem solver
Turning complexity into something simple enough to act on.
Team player
The best analysis is useless if the team can't use it.
Responsibility
Honest numbers, documented gaps, owned mistakes.
Off the clock
Coffee-powered late-night build sessions.
Portrait photography and aesthetic grooming.
Exploring Ahmedabad, usually via Rapido.
Webtoon reader — long-running Lookism fan.
Organising local community tech events.
Community & leadership
Helped organise the Startup Mini Meetup Ahmedabad, the Ahmedabad Innovation Summit and the FITAG Tech Expo 2026 — coordinating speakers, running logistics, keeping teams moving.
02
Toolkit
Technical depth · and the half that isn't technical
skills
📊
🐼🤖
Solid on the technical side — but none of it matters if I can't explain a finding to someone who never touches SQL. Both halves get equal attention.
C/03Getting Started with Artificial IntelligenceIBM SkillsBuild · Sep 2026
C/04Cloud Computing FundamentalsIBM SkillsBuild · Aug 2026
C/05Generative AI Essentials: Using LLMs to Work with DataIBM SkillsBuild · 2026
C/06Data Science Essentials with PythonCisco Networking Academy · 2026
C/07Data VisualizationKaggle · 2026
C/08PythonKaggle · 2025
C/09Elements of AIUniv. of Helsinki & MinnaLearn · 2025
C/10AI Tools & ChatGPTbe10X · 2025
Digital badges
Introduction to Large Language Models · IBMSummarizing Data Using IBM GraniteClassifying Data Using IBM GraniteGenerative AI Essentials · IBM SkillsBuildGoogle Analytics · GoogleGoogle Ads Measurement · Google
03
Selected work
Featured builds plus live data systems and field reporting tools below. Open a project to see the working application.
The problem: document Q&A tools mean uploading sensitive files to somebody else's server — a non-starter for regulated or confidential documents.
The build: a real retrieval-augmented Q&A pipeline that never leaves the browser — parsing, chunking, embeddings, semantic search and extractive answering, with no API keys and no cloud storage. Quantised MiniLM-L6-v2 for embeddings, DistilBERT-SQuAD for answer spans, both on-device via ONNX Runtime Web. Ships with role-based retrieval filtering and a live 6-question eval harness reporting recall, confidence and latency.
AI-powered threat detection & IOC analysis dashboard
The problem: security analysts watch indicators of compromise across five disconnected tools instead of one surface.
The build: a hybrid Python + React application for live monitoring — real-time feeds, automated IOC detection, ML-scored alerting and role-based dashboards for analysts and blue teams. Flask/FastAPI backend serving a Vite + React + TypeScript front end, SQLite for storage, and an analytics layer driving the live charts.
India's state-level education data — honest findings, null result included
The problem: state literacy and gender-gap data sits across government sources in inconsistent formats, with genuine holes (Telangana's 2014 split, missing exam records) that most write-ups quietly paper over.
The build: built on real 2015-16 UDISE school-census data, with a documented derivation for every computed metric and a pytest suite validating each number against an independent manual calculation. Gaps stay documented as gaps instead of zero-filled. I tested whether classroom crowding tracks exam pass rates expecting a clean story — it didn't (r ≈ −0.13), so it's published as a null result rather than reframed until something correlated.
Card-fraud decisioning, graded on the future and priced in euros
The problem: a fraud model has to decide in milliseconds on 0.17% prevalence, and then justify the decision. Almost every published model on this data is graded on a random split — trained on Tuesday, tested on Monday — which flatters it by 0.072 of average precision it will never see in production.
The build: an end-to-end system on 283,726 real transactions. Eleven models compared across expanding time windows, the threshold chosen by expected euro cost under a review-capacity constraint, SHAP reasons on every alert, PSI drift monitoring, and a FastAPI service that answers in 2.1 ms and provably reproduces the offline scores. 65 of 97 frauds caught for one false alarm, avoiding €7,071 of €11,596 exposure — on a window the model had never seen.
Free file converter — everything stays on your machine
The problem: every free PDF tool wants your file on their server first, for a task the browser can do by itself.
The build: merge, split, compress, rotate, watermark and AES-128 password-protect PDFs, plus conversion between PDF, Word, PowerPoint, Markdown, plain text and images — all client-side, no upload, no backend, no LibreOffice or Pandoc on a server anywhere. Built on pdf-lib, pdf.js, JSZip and mammoth with no framework or build step.
/01Delhi Air Quality 2021–20241,461 days of AQI readings — seasonality, pollutant relationships, publication-ready charts.Live ↗Code
/02Global GDP DashboardFlask + Plotly dashboard for GDP trends, forecasting and country comparison.Code ↗
/03COVID-19 Global AnalysisStreamlit dashboard — KPIs, filters and country-level comparison over Kaggle data.Code ↗
/04NetGuardNetwork intrusion detection — deep packet inspection across TCP/UDP/DNS/TLS with ML anomaly scoring.Code ↗
/05CryptoFlow-NotifyFull-stack blockchain alerting — MetaMask, Solidity contracts, Socket.io push for transfers and governance.Code ↗Story
/06SecureVaultLocal-only password manager — AES-256 via Fernet, PBKDF2 at 480k iterations, nothing leaves the machine.Code ↗
/07Ransomware Canary DetectorWatchdog-based early warning — hidden canary files in sensitive folders, instant alert on tamper.Code ↗
/08Sashakt Salon Impact DashboardLive impact reporting dashboard for salon beneficiary income, performance and operational KPIs.Live ↗Code
/09Sashakt Daily Activity LogDigital activity capture workflow for field teams, with structured daily reporting.Live ↗Code
/10Salon Beneficiary Income TrackerLive beneficiary income-entry system supporting structured field reporting.Live ↗Code
/11Live from GitHubFetching most recently pushed repositories…Profile ↗
04
Trajectory
Where this has been put to work
Experience
Jun 2026 — Present Current
Data & Digital Systems Analyst
Elixir Foundation · NGO, Ahmedabad
Own the beneficiary and programme data end to end — cleaning it, structuring it and analysing it so the team can see what's actually working on the ground and direct support where it lands hardest.
2026
Data Analyst Intern
BrainyBeam Technologies
Python, SQL, machine learning and Power BI turned into actionable insight for business logic decisions.
2026
CSR Social Impact Projects
Independent / volunteer
Ran beneficiary trend assessments, keeping the focus on whether the tech served the community rather than whether the dashboard looked impressive.
2024 — Present Ongoing
Data Analyst — Projects & Freelance
Self-employed, Ahmedabad
Independent dashboards, SQL analysis and Kaggle competitions — where most of my Pandas, Seaborn and query-optimisation instincts got built through repetition.
Education
2026 — Present Current
B.Tech, Information Technology
Aditya Silver Oak Institute of Technology
Going deeper into software systems, data engineering and the fundamentals sitting underneath the tools I already use daily.
2023 — 2026 · Completed
Diploma, Information Technology
Aditya Silver Oak Institute of Technology
Data structures, database management, statistics, machine learning, data mining and web development — the base the CSR work and my first real Python pipelines came out of.
Currently
Balancing the Elixir Foundation role with full-time study — which means every build here had to justify its own time.
Ruchir is an incredibly fast learner. Through our group projects he kept a positive attitude and was always open to taking feedback and acting on it. A true team player.
— Faculty recommendation
"
Organising the Tech Expo with Ruchir was seamless. Coordinating teams and running logistics while keeping everyone motivated is a rare kind of leadership.
— Event co-organiser
07
Contact
Let's build
something.
I'm available for small fixed-price freelance tasks and limited weekly project support — spreadsheet cleanup, dashboards, SQL, Google Sheets automation and reporting. I'm also happy to talk through a dataset you're stuck on. My inbox is genuinely open.
