Y 0000 · HERO
KR Production Trace Run 001
Record index Out ✓
Senior Java Backend Developer / 7+ yrs / Energy Market Settlement

Backend systems where correctness survives production.

I build Java backend systems for calculation-heavy domains: settlement logic, data validation, regression safety, production debugging, and complex enterprise workflows.

Open Trace Contact
Run ManifestVerified
Role
Senior Software Developer
Exp
7+ years backend
Stack
Java · Spring Boot · Oracle SQL
Markets
CAISO · EDAM · PJM · ERCOT
Base
Oklahoma · Remote
Status
Available · remote senior backend
017+ years backend development
02Mission-critical settlement systems
03Java · Spring Boot · Oracle SQL
04CAISO · EDAM · PJM · ERCOT
05Production debugging & regression safety
06Cross-functional delivery
Validation path — how a settlement result becomes trustworthy Move pointer to inspect stages
01 / Input
Input Data
ISO market files, meter and reference data
02 / Rules
Business Rules
Tariffs, charge codes, market behavior
03 / Calc
Settlement Engine
Charge-code logic turning rules into amounts
04 / Check
Validation
Checkpoints, reconciliation, regression tests
05 / Output · Verified
Settlement Output
Traceable, regression-guarded result
01 / Selected Work
4 records · select to inspect

Systems built, extended, and stabilized.

New market-region settlement behavior had to be implemented, stabilized, and proven correct against evolving rules — under transition deadlines.

Problem
Rules still moving; deadline fixed
Investigation
Trace output → calc → mapping → input
Validation
Expected behavior encoded as regression tests
Outcome
Behavior stabilized through go-live
Role
Primary developer / senior backend developer
Systems touched
Java · settlement engines · validation workflows
Validation strategy
Regression coverage on calculation logic; validation-support workflows confirming outputs against expected settlement behavior.
Tools
JavaJUnitOracle SQL
OutcomeSupported implementation and stabilization of EDAM-related settlement behavior.

Production calculation engines had to keep pace with changing tariffs and market rules without regressing existing charge behavior.

Problem
Tariff change without regression
Investigation
Charge-code behavior compared release to release
Validation
Regression tests guarding each charge type
Outcome
Engines extended, prior behavior intact
Role
Backend developer
Systems touched
Java · Oracle SQL · settlement calculations
Validation strategy
Extended engines with regression tests guarding greenhouse gas, grid management charge, forced outage, and EDAM-related behavior.
Tools
JavaOracle SQL
OutcomeMaintained and extended production calculation engines under evolving tariff and market rules.

Settlement runs surfaced discrepancies with no obvious single cause — input data, database mappings, and calculation logic all suspect.

Problem
Discrepancy with no single obvious cause
Investigation
End-to-end trace across every layer
Validation
Coverage added at the confirmed failure
Outcome
Faster resolution, higher confidence
Role
Senior backend engineer
Systems touched
Input files · calculation logic · database mappings · outputs
Validation strategy
Traced discrepancies end-to-end across inputs, logic, DB mappings, and outputs to isolate root cause and add coverage.
Tools
JavaOracle SQLLog tracing
OutcomeImproved issue resolution and validation confidence.

Settlement inputs and calculation results across multiple ISO market structures needed automated, repeatable accuracy checks.

Problem
Manual accuracy checks across ISO formats
Investigation
Input structures compared across PJM and ERCOT
Validation
Automated validation tooling over inputs and results
Outcome
Repeatable accuracy checks
Role
Software Developer
Systems touched
Java file processing · Oracle SQL · stored procedures · validation tools
Validation strategy
Built automated data-validation tools verifying settlement inputs and calculation results across ISO formats.
Tools
JavaOracleStored Procedures
OutcomeHelped ensure settlement input and calculation accuracy.
02 / Workbench

The Debugging Bench

Complex systems usually fail in the space between the rule, the data, and the assumption. My work is to trace that path, isolate the failure, and make the system safer for the next run.

Trace first. Guess later.
Suspected causesAwaiting trace
UNKNOWN
Input data
UNKNOWN
Database mappings
UNKNOWN
Existing charge logic
UNKNOWN
New charge-code path
Regression test pending
01
Understand the Rule
Start with the business rule, tariff change, or expected settlement behavior.
02
Trace the Data
Follow the input through files, database mappings, transformations, and calculation paths.
03
Isolate the Failure
Separate bad data, bad assumptions, mapping defects, and calculation-logic issues.
04
Write the Test
Turn the failure into regression coverage so the same bug does not return quietly.
05
Stabilize the Path
Fix the behavior, verify the output, and make the system easier to reason about.
06
Document the Behavior
Leave enough context for analysts, QA, operations, and future developers to understand the result.
03 / Capability Stack
6 layers · hover or focus a layer

Organized as an architecture, not a tag cloud.

Backend Systems
Java 8/11Spring BootHibernateJPAMavenLegacy Java
Used forBuilding and maintaining the calculation engines and enterprise workflows at the core of settlement systems.
Data & Persistence
OracleSQL ServerPostgreSQLStored proceduresLarge dataset extraction
Used forExtracting, transforming, and reconciling large settlement datasets and the mappings behind them.
Integration
REST APIsXML / JSONFile-based ETLISO market data files
Used forMoving ISO market data in and out reliably — files, APIs, and structured formats.
Reliability
JUnitMockitoRegression testingCode reviewsCoverage standards
Used forKeeping calculation logic correct across releases and catching regressions before production does.
Domain
ISO settlementsCAISOEDAMPJMERCOTEnergy market ops
Used forReading tariffs and market rules and turning them into settlement behavior that holds up.
Collaboration
AnalystsQAOperationsClient teamsDocumentationMentoring
Used forValidating results and resolving defects with the analysts, QA, ops, and client teams who depend on them.
04 / Experience
PCI Energy Solutions · Norman, OK

One system, growing more complex for seven years.

2015 — 2018
Software
Developer
M01Developed Java applications for processing ISO market data files and generating settlement reports.
M02Created SQL queries and stored procedures for extracting and transforming large datasets from Oracle databases.
M03Built automated data-validation tools to ensure accuracy of settlement inputs and calculation results.
M04Worked on PJM and ERCOT settlement modules, gaining experience across multiple ISO market structures.
JavaOracle SQLStored proceduresPJMERCOT
2018 — Present
Senior Software
Developer
M05Served as one of the primary developers for EDAM shadow settlement implementation and stabilization.
M06Focused on calculation logic, defect resolution, regression safety, and troubleshooting during a high-priority CAISO market transition.
M07Developed and maintained CAISO settlement calculation engines for greenhouse gas, grid management charge, forced outage, and EDAM-related settlement behavior.
M08Debugged production and validation issues in settlement runs; traced discrepancies across input data, calculation logic, database mappings, and settlement outputs.
M09Wrote comprehensive JUnit and regression tests for settlement logic; maintained code-coverage standards and helped prevent calculation regressions.
M10Implemented new charge-code logic based on CAISO tariff changes, regulatory requirements, and evolving market rules under tight deadline constraints.
M11Collaborated with business analysts, operations, client teams, and QA to validate settlement results and resolve calculation defects.
M12Participated in code reviews and mentored junior developers on Java best practices and settlement-domain knowledge.
M13Supported monthly settlement cycles by troubleshooting data-processing issues and ensuring timely completion of client deliverables.
JavaSpring BootOracle SQLJUnitSettlement enginesCAISO / EDAM
05 / About
Complex systems deserve receipts.

Kyle is a backend developer who works best in systems with real rules, real constraints, and real consequences.

His work has focused on settlement engines, data validation, production support, and calculation logic for energy market systems. He is strongest in domains where the answer is not just "write code," but understand the rule, trace the data, validate the result, and make the system safer for the next run.

He is seeking remote senior backend roles in financial services, energy platforms, enterprise backend systems, or other calculation-heavy software domains.

Trace complete · Output verified 06 / Contact

Available for remote senior backend roles.

For teams building calculation-heavy, data-sensitive, or production-critical backend systems.

Kyle Roller · Senior Backend DeveloperOklahoma · RemoteJava · Spring · OracleTrace 001 · Verified
KR / Production Trace hello@kyleroller.dev
A/01Full trace · sanitized
CAISOEDAMSettlement Engine

EDAM Shadow Settlement: Implementing and Stabilizing New Market Behavior

Serving as one of the primary developers implementing and stabilizing settlement calculation behavior for a new market region during a high-priority CAISO transition — where correctness had to be proven, not assumed.

Context

CAISO's Extended Day-Ahead Market (EDAM) introduced new settlement behavior that had to be represented in the calculation engine ahead of go-live. "Shadow settlement" means running the new logic in parallel against expected results before it drives real financial outcomes — a period where every discrepancy is a question that has to be answered, not waved through.

Problem

New charge-code and calculation logic had to be implemented against tariff rules that were still evolving, then stabilized so outputs matched expected settlement behavior — without regressing the existing charge types already running in production.

Constraints
Tariff and market rules still evolving during implementation.
Existing production charge types could not regress.
Shadow-settlement outputs had to match expected behavior before go-live.
Transition deadlines left little room for open questions.
Fig. 1 — Data flow: inputs resolve into a verified output
INPUT
ISO market & meter data
RULES
Tariff / charge-code logic
CALC
Settlement engine
CHECK
Validation & regression
OUTPUT ✓
Verified settlement
Role & System Area

Primary developer on the settlement calculation logic — the code that turns tariff rules and market inputs into charge amounts. Work spanned the calculation engine, the input/mapping layer that feeds it, and the regression suite that guards it.

Technical Decisions
Isolate new charge logic from proven paths
Kept EDAM-related charge-code logic separable so existing, verified calculations were not disturbed while the new behavior stabilized.
Encode expected behavior as regression tests
Turned expected shadow-settlement outcomes into JUnit and regression coverage — the spec for "correct" lived in tests, not memory.
Make discrepancies traceable end to end
Structured the work so any mismatch could be followed from input file to mapping to calculation to output, rather than inspected only at the result.
Document behavior for downstream teams
Left context analysts, QA, and operations could read — the result carried its reasoning with it.
Fig. 2 — Defect isolation: rule out layers until one remains
CLEARED
Input data
CLEARED
Database mappings
CLEARED
Existing charge logic
ISOLATED
New charge-code path
Inputs, mappings, and prior logic cleared against known-good references — narrowing the discrepancy to the new charge-code path, where it could be tested and fixed.
Debugging & Validation Approach
01Compare shadow-settlement output against expected results for each charge type.
02Trace each discrepancy from output back through calculation, mappings, and input.
03Separate data issues from mapping defects from calculation-logic issues before changing code.
04Add regression coverage for the confirmed failure, then fix and re-verify the path.
Fig. 3 — System state, before & after
Before
×New EDAM behavior not yet represented in the engine.
×"Correct" defined by expectation, not by tests.
×Discrepancies hard to attribute to a single layer.
After
EDAM charge logic implemented and matching expected results.
Behavior pinned by regression tests that fail loudly.
Discrepancies traceable input-to-output with documented cause.
Result

Supported the implementation and stabilization of EDAM-related settlement behavior through go-live. New charge logic matched expected shadow-settlement results, existing charge types stayed intact under regression coverage, and discrepancies had a documented trace rather than a guess behind them.

Tools Used
JavaOracle SQLJUnitRegression suiteSettlement engineLog tracing
What This Demonstrates
Comfort implementing calculation logic against rules that are still moving.
Turning "it should match" into regression coverage that keeps matching.
End-to-end tracing across data, mappings, and logic to isolate defects.
Stabilizing production-critical behavior under transition deadlines.
No proprietary internals, client data, or non-public logic disclosed.
Document modePrintable output of the trace

Kyle Roller

Senior Backend Developer
Oklahoma · Remote
linkedin.com/in/kyleroller
github.com/kyleroller
Professional Summary

Senior backend developer with 7+ years building calculation-heavy, mission-critical settlement systems for energy markets. Specializes in complex enterprise Java, settlement calculation logic, data validation, regression safety, and production debugging. Builds systems where correctness, traceability, and reliable outputs matter — and where business rules, data quality, and real-world consequences meet.

Core Competencies
BackendJava 8/11, Spring Boot, Hibernate, JPA, Maven
DataOracle, SQL Server, PostgreSQL, stored procedures
IntegrationREST APIs, XML/JSON, file-based ETL
ReliabilityJUnit, Mockito, regression testing, coverage
DomainISO settlements — CAISO, EDAM, PJM, ERCOT
PracticeProduction debugging, code review, mentoring
Professional Experience
Senior Software Developer · PCI Energy Solutions, Norman, OK
2018 — Present
Served as one of the primary developers for EDAM shadow settlement implementation and stabilization, focusing on calculation logic, defect resolution, regression safety, and troubleshooting during a high-priority CAISO market transition.
Developed and maintained CAISO settlement calculation engines for greenhouse gas, grid management charge, forced outage, and EDAM-related settlement behavior.
Debugged production and validation issues in settlement runs; traced discrepancies across input data, calculation logic, database mappings, and settlement outputs.
Wrote comprehensive JUnit and regression tests for settlement logic; maintained code-coverage standards and helped prevent calculation regressions.
Implemented new charge-code logic based on CAISO tariff changes, regulatory requirements, and evolving market rules under tight deadline constraints.
Collaborated with business analysts, operations, client teams, and QA/testing groups to validate settlement results and resolve calculation defects.
Participated in code reviews and mentored junior developers on Java best practices and settlement-domain knowledge.
Supported monthly settlement cycles by troubleshooting data-processing issues and ensuring timely completion of client deliverables.
Software Developer · PCI Energy Solutions, Norman, OK
2015 — 2018
Developed Java applications for processing ISO market data files and generating settlement reports.
Created SQL queries and stored procedures for extracting and transforming large datasets from Oracle databases.
Built automated data-validation tools to ensure accuracy of settlement inputs and calculation results.
Worked on PJM and ERCOT settlement modules, gaining experience across multiple ISO market structures.
Education
B.S. Computer Science
University of Oklahoma
Primary Stack
JavaSpring BootHibernateOracle SQLPostgreSQLJUnitMockitoMavenRESTETL
Kyle Roller — Senior Backend Developer Available for remote senior backend roles
KR / Production TraceRun 001
› initializing trace environment
› loading settlement ruleset · caiso / edam / pjm / ercot
› validation path active
› system ready