I build custom AI tools that turn your business data into decisions. Shipped in 2 weeks.
0
AI tools shipped
0 wk
average delivery
0k+
rows processed
0%
on-time delivery
About
I build data products that work in production.
Live tools with real users, not notebooks that live on my machine. My work spans the full stack: ingestion, cleaning, modeling, and user-facing analytics. What makes my background different is direct experience at the intersection of data and business, working on the Sales & Marketing side, which gives me the context to understand what a metric needs to drive before I write the first query.
Stack
Proficiency
ML Engineering
Platforms & Tools
Education
Computer Science Engineering (AI-ML)
New L J Institute of Engineering & Technology · Ahmedabad
4th Year2023 – Present
Engineering Plus: Python & Data Science
New LJ Institute · Certified
Completed2024
Experience
Marketing & Sales Associate
2026 – Present
AgilityRun end-to-end marketing and sales operations at an AI consulting firm, building and deploying automation systems for email, content, YouTube asset production, and outreach with Python, n8n, and make.com. Also handle structured B2B platform research and competitive analysis for global positioning.
Interests
Revenue YTD
$84.2k
01
Raw inputs turned into revenue-visible decisions.
02
Models trained on your data, deployed on your stack.
Interface
React · TypeScript · Tailwind
API Layer
Python · FastAPI · Auth
Database
PostgreSQL · Redis
Infra
Cloudflare · VPS · CI/CD
03
Frontend to database. Every layer owned and shipped.
Predictive Financial Analysis · 2025
Feed in your business numbers and see exactly where you're headed. ML model compares predicted vs. actual performance to surface trends and flag risks before they compound.
Instant Data Analysis · 2024
Upload any CSV and get a full column-by-column breakdown in seconds. Types, distributions, null rates, correlations. No code, no setup. Just answers.
AI-Powered Finance Tracker · 2026
Connect your real bank via Plaid. Automatic transaction sync, smart budget tracking, and AI-generated spending insights. Full-stack: React + FastAPI + PostgreSQL.
Loan Default Risk Engine · 2026
Classical ML through to a hybrid XGBoost + LSTM model, served behind a hardened FastAPI service with key-based auth. Calibrated probabilities, cost-based decision thresholds, fairness audits, and PSI drift monitoring.
Revenue
$84.2k
+12.4%
Customers
1,284
+8.1%
Accuracy
94.2%
+2.1%
Revenue vs Prediction
unique
1,247
null %
0.0%
Unique per row
unique
5
null %
1.2%
Top: Electronics (42%)
unique
890
null %
0.4%
$12 – $4,892 range
unique
365
null %
0.0%
Jan – Dec 2024
Sample Data
Total Balance
$12,847.43
+$2,400.00 this month
Recent Transactions
Netflix
Entertainment · 1h ago
-$15.99
Whole Foods
Groceries · 3h ago
-$87.42
Payroll
Income · Yesterday
+$4,500
Uber
Transport · Yesterday
-$12.30
above approve cut-off 0.18 · below decline 0.55
Top Risk Drivers · SHAP
Test AUC
0.665
PSI drift
0.04
Fairness gap
0.062
Internal tool · no public demo
Google Apps Script · 2026 – Present
Runs on my own inbox: auto-labels every thread on arrival, validates what is genuinely unread against what only looks unread, and writes a Sheets-backed dashboard that tracks how long replies actually take.
The result is no manual sorting, nothing missed, and real numbers on response speed instead of a guess.
Auto-applied labels
Median reply time
3h 12m
3–5 weeks
$5,000 – $10,000
1–2 weeks
$1,800 – $2,800
2–3 weeks
$3,200 – $6,000
1–2 weeks
$2,200 – $4,200