Sanyukta Suman
Full-Stack AI Engineer · Berlin Open to mid-level and senior-track roles

Best fit: AI Product Engineer · Data Engineer · Full-Stack Engineer

I build AI products, data systems, and workflow automation from prototype to production.

Co-founded Xonects and built the product architecture from the ground up. Previously owned analytics pipelines and reporting systems for 8+ e-commerce clients at Datadice, delivering measurable ROI.

Xonects

Shipped a cross-context AI system and real-time sync engine connecting five productivity apps in one place.

Built with Next.js, Express/Prisma, PostgreSQL, Redis, and Gemini AI. Backed by NVIDIA Inception, Microsoft for Startups, and ZEROBASE.

xonects.com — your conversational work OS
Xonects — AI that understands context across all your apps and executes for you

Co-founder · AI Work OS · Nov 2025 – Present

Backed by NVIDIA Inception Program · Microsoft for Startups · ZEROBASE

What I owned

Architected the multi-app OAuth integration layer. Built the real-time sync protocol between five apps. Designed the Gemini-powered prompt chain that retains session memory across tools.

Scale

Connects Gmail, Outlook, Notion, Trello, and Google Calendar through a unified sync layer, enabling cross-app AI execution.

Impact

Live product with working demo. Architecture supports cross-app context retention and multi-platform AI execution — teams stop context-switching.

Next.js TypeScript PostgreSQL Gemini AI Docker
Datadice

Replaced weekly manual exports with real-time client dashboards.

Built the ingestion-to-visualization pipeline for 8+ e-commerce clients at Datadice. BigQuery ETL processed millions of daily transactions from Shopify, Amazon, and ad platforms. Result: 20%+ ROI lift for clients and 50% faster reporting.

Looker Studio — live client dashboards
Datadice — social media marketing dashboard

Data Analyst · Datadice GmbH · Jun 2022 – Aug 2025

What I owned

Designed schemas, modeled transformations in dbt, and maintained client-specific Looker Studio views for 8+ accounts end to end.

Scale

Millions of daily events from Shopify, Amazon, Google Ads, Facebook, and Instagram — processed through BigQuery pipelines with automated quality checks.

Impact

Clients reallocated ad budgets based on data-driven insights, driving 20%+ ROI improvements. Internal reporting dropped from weekly manual exports to daily auto-refresh.

BigQuery Looker Studio dbt Python
Folosolo

Built a multi-touch attribution system that gives marketing teams a defensible way to assign credit and budget.

Uses Shapley values to fairly distribute conversion credit across every channel, replacing last-touch guesswork. Spans clickstream simulation (Airflow), data modeling (BigQuery + Dataform), ML attribution (Random Forest + SHAP), and a live React dashboard.

folosolo.com — FOLOSOLO attribution dashboard
FOLOSOLO dashboard

Solo Project · Marketing Attribution · folosolo.com

What I owned

Wrote Airflow DAGs that generate synthetic multi-touch journeys. Built Dataform models to prep journey-level training data. Trained a Random Forest and extracted Shapley values per channel. Built the React visualization layer for per-channel ROAS.

Scale

Pipeline runs entirely on cloud infrastructure — Airflow orchestrates Python jobs, BigQuery stores raw and modeled data, and the React dashboard queries precomputed Shapley scores.

Impact

Replaces last-touch attribution with a math-based model. Full documentation and live demo at folosolo.com.

Python scikit-learn / SHAP BigQuery / Dataform Apache Airflow
Unwritten Data

Upload a CSV. Describe your brief. Get a client-ready analysis in minutes.

AI-powered analysis that turns any CSV into a structured report with statistics, charts, and AI recommendations — no code required. Built for consultants who need fast, defensible insights.

data-consultant.vercel.app — AI-powered data analysis
Unwritten Data — results dashboard with charts and KPIs
Data profiling and AI insights AI-generated recommendations Next steps and action items

Solo Project · Data Analysis App · Unwritten Data

What I owned

Built the complete app: Next.js frontend with Recharts, Supabase/PostgreSQL for persistence, and Gemini AI for structured insight generation.

Scale

Handles arbitrary CSV uploads, runs server-side statistics and AI inference, and returns structured reports in real time.

Impact

Consultants go from raw CSV to client-ready analysis in minutes — demonstrating product thinking and AI product execution.

Next.js 16 TypeScript Supabase / PostgreSQL Gemini AI Recharts
What People Say

Testimonials

Sanyukta built our entire BigQuery pipeline from scratch — processing millions of transactions daily — and delivered Looker Studio dashboards that cut our reporting time in half. The 20% ROI lift she drove was measured and directly tied to her work.

Thomas Knorr

CEO · Datadice GmbH

We came to Sanyukta with a messy data problem — scattered offers, no pipeline, no ML. She built the entire recommendation engine from scratch and our user engagement jumped 15% within months.

Desh Deepak

CEO · GottData

Sanyukta can take a fuzzy product idea and turn it into a working system within days. The cross-context AI architecture she designed is genuinely innovative.

Harit Krishan

Collaborator · Xonects

Where I've Worked

Experience

Datadice

Datadice GmbH

Data Analyst

Jun 2022 – Aug 2025 · Coburg

Promoted from Junior Data Analyst to a higher-ownership role over 3 years. Owned ingestion, modeling, and visualization pipelines across 8+ e-commerce clients.

BigQuery Looker Studio dbt
GottData

GottData

Freelance Data Science Consultant

Mar 2023 – Dec 2023 · Remote

Built a recommendation engine for health insurance offers that increased user engagement by 15% across 10,000+ monthly active users. Developed ML pipelines predicting lead conversion for an Icelandic marketing client, boosting conversion rates by 20%. Freelance, concurrent with Datadice.

BigQuery ML Vertex AI scikit-learn AWS RDS A/B Testing
UniCredit

UniCredit

Data Engineering Intern

Mar 2021 – Aug 2021 · Munich

Built ETL data pipelines for risk analytics on large-scale financial datasets. Collaborated with senior engineers on data quality validation and automated reporting for regulatory submissions.

Python SQL ETL Pipelines
What I Can Own

Capabilities

Product & Frontend

Next.js / React TypeScript Tailwind CSS

Backend & Infrastructure

Node.js / Express PostgreSQL Redis Docker GCP / AWS

Data & ML

BigQuery Looker Studio SQL Python Gemini / Vertex AI scikit-learn
Academic Background

Education

M.Sc. Data Engineering

Jacobs University, Bremen

2019 – 2021

B.E. Computer Science

BMSIT, Bangalore

2014 – 2019
Get In Touch

Let's talk.

Open to mid-level and senior-track roles in AI product engineering, data engineering, and full-stack product development in Berlin or remote. Best way to reach me: email or LinkedIn.