Portrait of Vijay M. Surve

Get out of your
comfort zone.

Scientific Software Developer & Data Systems Analyst

Vijay M.
Surve

I build Python/SQL data pipelines, ML-driven analytical workflows, and BI dashboards for operational and scientific environments — currently embedded at Fermilab.

USA Fermilab Open to LCLS & beamline data challenges
01

About

Scientific software developer and data systems analyst with hands-on experience building Python/SQL data pipelines, ML-based analytical workflows, and BI dashboards in operational environments.

Currently embedded at Fermilab, supporting inventory data systems and computational pipelines for scientific facility operations. Familiar with LLM integration, REST APIs, and workflow automation. Eager to bring data engineering and software development skills to LCLS beamline data challenges.

Languages

Python (Pandas, NumPy, scikit-learn, REST APIs), SQL (joins, CTEs, window functions), R, VBA

ML / AI

Scikit-learn, LLM integration (Claude API, prompt engineering, structured outputs), ML forecasting models

Data & Pipelines

Azure Databricks, Azure Synapse, Oracle EBS, SQL Server; ETL design, data validation, lineage

BI & Reporting

Power BI, Tableau; KPI dashboards, ad hoc packs, stakeholder reporting; familiar with MicroStrategy

Automation

Power Automate, REST integrations, Python scripting, Excel/VBA macros

Education

MS, Business Analytics — Clark University (3.8 GPA)
BE, Electronics Engineering — University of Mumbai

02

Experience

Sep 2024 — Present

Database Administrator — Business Operations & Reporting

Fermilab, Lead, SD

  • Developed Python/SQL pipelines to unify WMS and BOL data streams across scientific facility systems; applied validation checks reducing data issues by 25%.
  • Built Power BI dashboards surfacing inventory and throughput KPIs used by operations stakeholders for daily and ad hoc decision-making.
  • Standardized 300+ component records using Excel/VBA; improved data classification consistency across the component tracking system (CDB).
  • Automated report delivery via Power Automate; eliminated 1,000+ manual data entries per month, improving accuracy and freeing analyst capacity.
  • Gathered requirements with end users and scientists; documented end-to-end workflows and system handoffs across multiple operational platforms.
  • Led UAT and release testing for new reporting flows; logged defects, verified fixes, and maintained release notes for iterative deployments.
  • Maintained implementation tracker and data dictionary; improved data lineage visibility and onboarding speed for new team members.
Jun 2023 — Jan 2024

Business Intern — Financial Systems & Analytics

Fermilab, Batavia, IL

  • Automated BPS to Oracle EBS reporting workflows using Python and R; accelerated close cycle insights by 40%.
  • Built Tableau KPI dashboards aligned with stakeholder metric definitions; added QA validation checks to ensure data integrity.
  • Trained teams on updated dashboards and analytical tools; reduced support tickets and improved adoption across departments.
  • Documented SAP Concur approval workflows; authored user guides to improve process consistency and supportability.
Dec 2021 — Jul 2022

Junior Business Analyst

Vagmine Marine Care, Mumbai, India

  • Built ML-based demand forecasting models; delivered $100K+ in inventory savings and 20% YoY revenue improvement.
  • Designed Oracle-to-Azure Synapse ETL pipeline; improved processing speed by 25% with validation checkpoints.
  • Cleaned and restructured finance datasets using SQL and Access; improved audit accuracy and reporting trust.

Awards & Recognition

2025

EPRA Award & R&R Award — Fermilab

2026

Director's EPRA Award — Fermilab

03

Projects

2024

LLM-Integrated Scientific Data Query Tool

A Python application integrating the Claude API with structured scientific inventory data, enabling natural language querying of component lifecycle and location history.

  • Python
  • Claude API
  • Prompt Engineering

Designed prompt engineering workflows to extract structured outputs from unstructured lab notes — demonstrating LLM applicability to scientific operations.

2024

FinStream — Real-Time Financial Analytics Pipeline

A real-time data pipeline on Azure Databricks achieving sub-60 second end-to-end latency from ingestion to dashboard.

  • Azure Databricks
  • Python
  • SQL
  • scikit-learn

Applied ML models for signal detection and anomaly flagging, backtested on 10 years of market data via the Polygon.io API.

04

Connect

Have a data challenge, a beamline problem, or just want to talk pipelines? Reach out — I'd love to hear from you.