Jashan.
Jashan Sadioura

Forward-Deployed Engineer · AI Product Manager

I build AI agents that enterprise data teams trust in production.

Hi, I'm Jashan Sadioura. I embed with enterprise data teams and turn their slowest manual work into governed AI agents people actually trust. My latest cut data-change reviews from a day or more to under an hour.

I don't treat AI as a chatbot problem. I treat it as a product, data and governance problem, which is why what I ship stays in production.

View my workDownload resume

Open to FDE and AI PM roles

Gurugram, India
Based in
6.5+ years in data & AI
Experience
AI, Data, Product & Governance
Focus areas
Data Radian TechnologiesDeloitteSmart Energy WaterMBA — Thapar Institute of Engineering and TechnologyB.E. Computer Science — ThaparIEEE published · cited 10+Microsoft Certified: AI Product Manager
02About

From analytics to AI systems in production

I'm a Forward-Deployed Engineer and AI Product Manager. I work inside enterprise data teams, find the repetitive, high-stakes work, and turn it into AI systems people trust enough to use every day.

How I got here

  1. 2020–2022

    Product Analyst → Senior Product Analyst

    Smart Energy Water · B2B SaaS

  2. 2022–2024

    Analytics Specialist

    Deloitte

  3. 2024–2026

    Data Quality Analyst Lead

    Data Radian Technologies

  4. 2026–present

    Forward-Deployed Engineer

    Data Radian Technologies

How I work

01

Trust before dashboards

A dashboard is only worth building once people trust the numbers in it. I fix quality, ownership and lineage first, then ship the surface.

02

Code for answers, models for judgement

If a step has a testable correct answer, it's deterministic code. I spend a model only where correctness is a judgement call, and I measure it before I believe it.

03

Humans own the decision

My agents gather evidence and say what they found. A person owns the merge and the sign-off. A wrong automated pass costs more than a slow human one.

03Work

Selected work

01 · Case study

An agentic QA system that reviews data changes before they ship

I designed and shipped a five-dimension quality review that runs against proposed data changes: deterministic Python owns everything with a testable correct answer, bounded LLM workers own only the judgement calls, and every result is advisory because a human reviewer owns the merge decision.

  • Review: 1+ day → under 1 hour
  • Running in production
  • Same evidence, any analyst
  • Python
  • LLM agents
  • Pydantic contracts
  • Snowflake (read-only)

02 · Case study

From ad hoc testing to a 2-day QA SLA

I replaced ad hoc, single-analyst-dependent QA testing at Fortune Brands with a standardized process any analyst could run — cutting ticket turnaround to a 2-day SLA with full evidence documentation.

  • 2-day QA ticket SLA
  • 1000s of tests → Snowflake DMFs
  • Any analyst covers any ticket
  • Snowflake
  • dbt
  • SQL
  • Tableau

03 · Case study

A payment score that lifted on-time payments 20–30%

I built a payment risk score for Smart Energy Water's retail customers that let the business target at-risk payers directly — driving a 20–30% increase in on-time payments.

  • +20–30% on-time payments
  • Scoped to retail payers
  • Power BI
  • SQL
  • Product discovery

04 · Case study

dbt-sentinel: measuring a review agent before believing it

I built dbt-sentinel, an agent that reviews dbt pull requests by computing the blast radius of a diff — and published its eval numbers including the targets it missed.

  • 30 labelled fixtures
  • Eval-gated CI
  • Misses published
  • Python
  • Graph traversal
  • LLM (gpt-4o)
  • Evals
GitHub ↗
04Experience

6.5+ years across product, analytics and data

Data Radian Technologies

Client: Fortune Brands Innovations

Gurugram, India

  • Forward-Deployed EngineerJul 2026 — Present
  • Data Quality Analyst LeadApr 2024 — Jul 2026

2 roles since Apr 2024

Show details

Forward-Deployed Engineer

Jul 2026 — Present
  • I designed and shipped a multi-agent QA system that reviews proposed dbt and Snowflake data changes across five quality dimensions. A coordinator agent plans each run, deterministic Python owns every check with a testable answer, and bounded LLM workers handle only the judgement calls.
  • I cut data-change review time from a day or more to under one hour. It runs in production in the team's review workflow, and the same change produces the same evidence no matter who runs it.
  • I built governance into the design: a read-only warehouse role, privacy-by-design evidence (aggregates and masked values, never raw rows), and advisory output, so a human reviewer owns every merge decision.
  • I built, tested and validated a Snowflake semantic model and an AI assistant on Snowflake Cortex on top of it.
  • I'm building AI assistants for a reporting platform migration onto certified data using Snowflake streams (in progress).

Moved from leading data quality reviews to building the AI system that automates them.

Data Quality Analyst Lead

Apr 2024 — Jul 2026
  • I managed 6–7 data quality analysts validating Data Vault pipelines on Snowflake and dbt end to end, from ingestion through transformation to the reporting layer, via GitHub pull request reviews and root cause analysis.
  • I standardised the QA operating model (shared scripts, testing templates, documentation and approval process) without disrupting delivery, established a 2-day SLA for QA tickets, and cross-trained analysts across engagements.
  • I introduced regression testing where none existed: I built the framework in Snowflake Data Metric Functions, migrated thousands of tests from dbt, and automated weekly monitoring emails.
  • I established a data governance framework covering PII identification, source documentation, data ownership mapping and usage governance.
  • I built a centralised Tableau data quality product for monitoring production data quality metrics and regression outcomes.
Deloitte

Gurugram, India

  • Analytics Specialist (AERS)May 2022 — March 2024
Show details

Analytics Specialist (AERS)

May 2022 — March 2024
  • I owned end-to-end analytics delivery for Audit & Assurance engagements, working with large-scale financial datasets for reconciliation, validation, and audit analytics.
  • I acted as analytics product owner for multiple audit use cases, translating requirements into scalable Tableau dashboards, VBA automation tools, and standardized analytics workflows.
  • I designed and maintained custom Tableau dashboards across audit domains — revenue analytics, contract assets, unbilled revenue, click-rate-based revenue, and shipments & flights.

    Improved audit efficiency and enabled full-population analysis instead of sample-based testing; surfaced anomalies and revenue-leakage patterns through interactive visual analytics.

  • I built a VBA-based Trial Balance automation tool that ingests raw trial balance and chart of accounts data and generates a fully formatted, audit-ready output in seconds.

    Reduced audit preparation time by ~90–95% — from multiple days to seconds — significantly improving engagement turnaround and analyst productivity.

  • I led Tableau enablement and analytics upskilling initiatives, training new hires and increasing self-service analytics adoption across the practice.
Smart Energy Water

Gurugram, India

  • Senior Product Analyst / Product AnalystJan 2020 — Apr 2022
Show details

Senior Product Analyst / Product Analyst

Jan 2020 — Apr 2022
  • I led the full SDLC for multiple product initiatives — requirements discovery, PRDs/FRDs, user stories, UX collaboration, sprint execution, and UAT.
  • As a core contributor to the Customer Analytics team, I designed and delivered Customer Journey Analytics in Power BI, providing visibility into app usage trends, drop-off points, and high-friction pages.

    Insights led to targeted UX and feature optimizations, increasing user engagement, session time, and overall app adoption.

  • I built Payment Score Analytics to classify customers by payment behavior — on-time, delayed, defaulting — and segment payment risk.

    Enabled business teams to target high-value customers more effectively and design tailored plans for at-risk users, improving collection rates and retention.

  • I drove the successful integration of digital payment services, aligning product capabilities with client and end-user needs.
05Capabilities

Skills, grouped by the problem they solve

01

AI product strategy

  • AI use-case identification
  • Discovery & user research
  • PRDs & roadmaps
  • Experimentation
  • Stakeholder alignment
02

AI systems

  • AI agents
  • LLM workflows
  • Evaluation & human-in-the-loop review
  • Test automation
03

Enterprise data products

  • Data quality frameworks
  • Governance policy
  • Lineage & observability
  • DPDP / GDPR awareness
  • Snowflake
  • dbt
  • SQL
  • Python
  • Data modeling
04

Product analytics

  • Metric design
  • Behavioural analytics
  • Executive reporting
  • Tableau
  • Power BI
  • Excel analytics
06Building

Outside the day job

Data Literacy for Leaders

A YouTube channel teaching non-technical executives to read, question, and govern their own data — covering governance, privacy compliance, and AI bias.

Watch on YouTube

Writing on AI & data products

Long-form notes on the work itself — AI product management, data strategy, data quality, and enterprise AI.

Read the blog