Skip to content
Founder @ Exactis

Adam Jalal

Data Engineer, Cloud Engineer, Founder @ Exactis

I build data systems you can trust: ETL pipelines, real-time streaming, cloud infrastructure and MLOps.

Paris · Dublin

View my work Download Resume
View resume online ↗

01 About

Data that moves.
Data you can trust.

Data and Cloud Engineering student on a dual-degree programme (EPITA / Griffith College), specialised in building ETL pipelines, vector search (FAISS) and MLOps deployment. Founder of Exactis, a B2B SaaS reconciling industrial production data against ERP systems.

Languages

  • FR

    French

    Native

  • IT

    Italian

    Native

  • EN

    English

    C1 / Fluent

  • AR

    Arabic

    Fluent Spoken

What I focus on

  • Data pipelines

    01

    ETL, orchestration and data quality, from raw multi-source data to KPIs people can trust.

  • Cloud & MLOps

    02

    Containers, Kubernetes, autoscaling and observability to run models and services in production.

  • Reliable reconciliation

    03

    Matching production data against ERP references with zero silent errors, the core of Exactis.

  • AI-assisted engineering

    04

    Shipping faster and cleaner with Claude Code, Claude, Gemini and Google AI Studio.

Currently

  • Founder @ Exactis
  • Year 3 @ Griffith College, Dublin
  • EPITA, Paris
Portrait of Adam Jalal
Adam JalalData & Cloud Engineer · Founder @ Exactis
~/adam/profile.json
{
"name": "Adam Jalal","role": "Data & Cloud Engineer","company": "Exactis","based": ["Paris", "Dublin"],"focus": ["ETL", "Streaming", "Cloud", "MLOps"],"ai_tooling": ["Claude Code", "Gemini", "MCP"]}
Paris · DublinUTF-8 · JSON

02 Impact in numbers

Results, measured.

  • 9.17M+

    records processed

    NYC Taxi ETL

  • -82%

    PostgreSQL load time

    NYC Taxi ETL

  • 1h+/day

    manual work eliminated

    Lesieur Cristal

  • 88%

    explained variance (R²)

    AptiSkills

  • 1,279

    vectors, sub-second search

    AptiSkills · FAISS

  • -60%

    Docker image size

    CloudPredict

03 Founder

Exactis.

B2B SaaS reconciling industrial production data against ERP systems.

Founder · Paris, France · Sept 2026 - Present

  • 01

    Building an AI-assisted B2B SaaS that reconciles factory production data against ERP (SAP) reference data, replacing manual Excel cleaning in industrial SQDC reporting.

  • 02

    Designed and deployed the reconciliation engine (Vercel serverless): Jaro-Winkler similarity ranks candidate matches while exact-match-only validation guarantees zero silent errors, with near-matches routed to human review.

  • 03

    Diagnosed that pure similarity scoring silently auto-validated transposed product codes (98.3% match on a wrong reference) and re-architected the decision logic to close it; covered by a regression test suite.

exactis.pro adam@exactis.pro
reconciliation-engine · run #042
exact matchnear-match
Production · MESERP · SAP reference
  1. 3021-0112-403021-0112-40 Exact match - auto-validated
  2. 3021-0458-723021-0458-27 Transposed code - blocked→ human review
  3. 3021-0387-153021-0387-15 Exact match - auto-validated
  4. 3021-0520-093021-0520-09 Exact match - auto-validated

3

validated

1

human review

0

silent errors

Illustrative data. Similarity scoring alone would auto-validate 3021-0458-72 against 3021-0458-27 (98.3% Jaro-Winkler). Exactis only auto-validates exact matches and routes near-matches to a human.

04 Experience

Where I've shipped
real data work.

  1. Sept 2026 - Present

    Paris, France

    Current
    • Building an AI-assisted B2B SaaS that reconciles factory production data against ERP (SAP) reference data, replacing manual Excel cleaning in industrial SQDC reporting.
    • Designed and deployed the reconciliation engine (Vercel serverless): Jaro-Winkler similarity ranks candidate matches while exact-match-only validation guarantees zero silent errors, with near-matches routed to human review.
    • Diagnosed that pure similarity scoring silently auto-validated transposed product codes (98.3% match on a wrong reference) and re-architected the decision logic to close it; covered by a regression test suite.
    • Vercel
    • SAP
    • Jaro-Winkler

  2. Apr 2026 - Jul 2026

    Casablanca, Morocco

    Data Engineer Intern

    Lesieur Cristal

    • Architected a multi-source data pipeline (Evocon MES API, SAP ERP via MB51, Excel) automating the weekly SQDC performance review and eliminating over 1 hour/day of manual Excel compilation.
    • Diagnosed and resolved a critical DAX aggregation bug (an erroneous 2,054% KPI) by redesigning a Power BI star-schema model (Fact_Arrets + 3 dimension tables) with DIVIDE()-based measures, delivering mathematically accurate reporting.
    • Authored a Business Requirement Document for Group Avril's IT department to formalise SAP data access (movement codes 101/102/261/262), targeting >95% data reliability and full automation of the performance canvas.
    • Power BI
    • DAX
    • SAP
    • Evocon MES API
    • Excel

  3. May 2025 - Jun 2025

    Levallois-Perret, France

    R&D Data Engineer Intern

    AptiSkills

    • Cleaned and structured 74 composite materials (2,560 raw rows) from the Sandia MHK database into a relational SQLite model (Matériau/Résine/Essai).
    • Built a scikit-learn regression pipeline predicting maximum stress with 88% explained variance (R²).
    • Implemented an L2 FAISS vector index for sub-second similarity search across 1,279 material vectors, cutting manual lookup time from hours to minutes, and delivered it via an end-to-end Streamlit app.
    • Python
    • SQLite
    • Scikit-learn
    • FAISS
    • Streamlit
    • Pandas

05 Selected projects

Pipelines, platforms
& search engines.

All repositories

  • NYC Taxi ETL

    9.17M+ records, orchestrated end-to-end, and loaded 82% faster.

    NYC taxi records processed
    9.17M+
    NYC taxi records processed
    PostgreSQL load time (~51 → ~9 min)
    -82%
    PostgreSQL load time (~51 → ~9 min)
    • Python
    • Pandas
    • PostgreSQL
    • Prefect 3
    • SQL
    • SQLAlchemy
    • +1
    Read case study
  • 2-person team

    CloudPredict

    A cloud-native MLOps platform that scales itself from 2 to 10 pods.

    Docker image size (multi-stage)
    -60%
    Docker image size (multi-stage)
    pods, autoscaled at 50% CPU
    2 → 10
    pods, autoscaled at 50% CPU
    • Docker
    • Kubernetes
    • HPA
    • RBAC
    • Nginx Ingress
    • Prometheus
    • +4
    Read case study
  • 2-person team

    Real-Time Shop Data Streaming

    Sales tickets streamed through Kafka, stored, monitored and exposed live.

    sales pipeline on a Kafka backbone
    Real-time
    sales pipeline on a Kafka backbone
    stock threshold triggering alerts
    <10%
    stock threshold triggering alerts
    • Kafka
    • PostgreSQL
    • Docker Compose
    • Prometheus
    • Grafana
    • Python
    Read case study
  • Built during my AptiSkills internship

    Materials Similarity Search

    From scattered lab spreadsheets to sub-second similarity search.

    composite materials structured (2,560 raw rows)
    74
    composite materials structured (2,560 raw rows)
    explained variance (R²) on max stress
    88%
    explained variance (R²) on max stress
    • Python
    • SQLite
    • Pandas
    • Scikit-learn
    • FAISS
    • Streamlit
    Read case study

06 Skills

The stack behind
the pipelines.

tip: hover a technology to light up where I used it.

Programming & Scripting

4 tools

Data & Cloud

11 tools

AI, ML & Visualisation

10 tools

AI-Assisted Engineering

6 tools

Highlights the experiences and projects that use this technology.

07 Education

Two schools,
one dual degree.

  • Sept 2026 - Jun 2027

    Griffith College

    Dublin, Ireland

    BSc (Hons) in Computing Science - Year 3, Dual Degree

    Dual Degree

    Key modules

    • Linux System Administration
    • Concurrent Development
    • Relational Databases
    • Networks & Data Communications
    • Data Structures & Algorithms
    • Web Technologies
  • Sept 2024 - Jun 2027

    EPITA

    Paris, France

    Programme Grande École

    Specialisation

    • Cloud Engineering
    • Machine Learning
    • Advanced Algorithms
    • Linear Algebra
    • Systems Architecture

08 Certifications

Certified, beyond the code.

  • 02

    AI & Law

    Lund University

    View certificate
  • 03

    GDPR Foundations

    Measured Collective

    View certificate
  • 04
    Certificate of participation

    Ethics and Governance of Artificial Intelligence for Health

    World Health Organization

09 Leadership & Involvement

Beyond the terminal.

  • Nov 2025 - Jun 2026Paris

    Student Ambassador

    EPITA

    Represented EPITA's AI & Management programmes to prospective students at key events, driving engagement and academic visibility.

  • May 2025 - Sept 2026Paris

    Activist

    Amnesty International France

    Supported human rights awareness campaigns and citizen mobilisation efforts over a 16-month period.

  • Jan 2025Paris

    Member, “Carré du Dean”

    ISG - Institut Supérieur de Gestion

    Co-authored strategic business reports on Toyota, LVMH and OpenAI; drafted an international-partnership MoU.

  • May 2025Paris

    Startup Immersion (Shadowing)

    Soro, Station F - world's largest startup incubator

    Observed DeepTech growth dynamics and Agile methodologies firsthand within the Station F ecosystem.

10 Contact

Let's buildsomething reliable.

Internships, missions, freelance work, or simply a question: my inbox is open.

  • Internships
  • Missions
  • Freelance
  • Questions
Write to me
Call
LinkedIn/in/adam-jalal-techGitHub@AdamJalalIA
Download Resume View online

About Exactis specifically? Reach me at adam@exactis.pro or visit exactis.pro.

Copied