Hi, I'm

Roshan Ali

Senior AI QA Engineer  ·  ERP & AI Platform QA  ·  API & Data  ·  Agile Shift-Left QA

Download CV View Projects
👨‍💻
Scroll
Professional Summary

5+ Years of Quality Excellence

90+
Agile Sprints Delivered
1,200+
Defects Governed to Closure
2,500+
Test Cases Executed
67+
Tech Specs Reviewed

Engineering Quality in the Age of Intelligent Systems

ISTQB Certified Test Manager (CTAL-TM) and Lean Six Sigma Green Belt with 5+ years of hands-on experience architecting quality strategies for complex, non-deterministic ERP and AI scheduling platforms. Proven track record of orchestrating end-to-end validation across functional, performance, and security-aware testing while remaining deeply embedded in API, data, and system-level verification.

Drove zero-critical-defect releases by pioneering AI-assisted testing, optimising regression coverage, and accelerating root-cause resolution across sophisticated software-hardware implementations. Specialised in designing custom constraint-driven frameworks — rules, invariants, feasibility, cost, and impact checks — for non-deterministic AI output validation.

Recognised for taking full ownership of quality outcomes, mentoring senior QA engineers, and influencing cross-functional teams to adopt shift-left practices that prevent defects before they reach development.

ISTQB CTAL-TM Lean Six Sigma Green Belt Risk-Based Testing Shift-Left QA Non-Deterministic AI Validation ERP Platform QA Defect Governance API & Data Validation
Core Competencies

Tools & Expertise

A comprehensive toolkit for end-to-end quality engineering across modern enterprise environments.

Test Management & Strategy

Test Strategy & PlanningRisk-Based Testing Release ReadinessJIRA XrayConfluence TestRailRTC

ERP & AI Platform QA

E2E ERP TestingAI PSO Scheduling Non-Deterministic Validation Constraint-Driven Frameworks Invariant ChecksFeasibility Testing

API & Data Validation

PostmanSwagger SQL / Database Validation Data IntegrityLog Analysis End-to-End Data Correctness

Functional & Regression Testing

Functional TestingRegression Testing Integration TestingSmoke Testing UATProduction Sign-Off Exploratory Testing

Defect Lifecycle Governance

Defect TriageRoot-Cause Analysis Severity DisciplineQuality Gate Reviews Reproducibility StandardsDefect Aging

Delivery & Agile QA

Agile / ScrumShift-Left QA Jenkins CI/CDSprint Planning Tech Spec ReviewRelease Risk Reporting

Test Automation

SeleniumAI-Assisted Test Design Regression OptimisationDefect Clustering JAVAAutomation Frameworks

Leadership & Communication

Stakeholder ReportingCross-Functional Alignment QA MentorshipQuality Gate Ownership Go/No-Go DecisionsDashboard Ownership
Career Journey

Experience Timeline

A decade of quality ownership and end-to-end strategy for enterprise-level products.

Apr 2022 — Dec 2025
Senior Software Engineer – Quality Assurance
IFS  (IFS R&D International Pvt Ltd)

Architected end-to-end quality strategy for a non-deterministic ERP & AI PSO scheduling platform — single-handedly owning test strategy, risk-based prioritisation, and release readiness across 90+ sprints, consistently delivering on-time, high-confidence releases. Engineered and executed 2,500+ test cases across 510+ user stories spanning functional, regression, integration, smoke, and UAT, achieving full traceability from specs to outcomes with zero escapes to production.

Drove defect governance at enterprise scale: triaged and accelerated resolution for 1,200+ defects, enforcing reproducibility standards and severity discipline that cut average defect age by 35%. Led 80+ defect triage and quality gate reviews as the sole quality lead, aligning Dev, Product, and stakeholders. Pioneered constraint-driven validation for non-deterministic AI outputs — rules, invariants, feasibility, cost, and impact checks — that became the team's standard methodology. Mentored and upskilled 3 senior QAs, reducing review rework by 25%, and championed AI-assisted QA adoption across R&D.

ERP & AI Platform QANon-Deterministic Validation Postman + SQLJira & Xray Shift-Left QARelease Readiness QA Mentorship
Oct 2021 — Apr 2022
Industrial Engineer
MAS Holdings (Thurulie)

Overhauled data collection and reporting processes, eliminating manual inconsistencies by 30% and boosting operational data integrity by 25% across tracking and reconciliation systems — directly improving decision-making accuracy. Established 10+ operational metrics and acceptance criteria for process changes, enforcing outcome validation before every rollout and maintaining 100% Integrated Management System compliance across health, safety, quality, and environmental standards.

Data IntegrityProcess Validation Operational MetricsIMS Compliance Acceptance Criteria
Work Gallery

Featured Projects & Artifacts

Test strategies, frameworks, and QA artifacts from real enterprise engagements.

01Optimization EngineEnterprise Scheduling Optimization Engine
Business Goal
Owned quality for a constraint-aware scheduling engine built to maximize workforce productivity, protect service commitments, reduce avoidable travel, and keep operations scalable as workload increased.
System / Module
Core optimization layer: activity-to-resource assignment, continuous plan improvement, large search spaces, travel reduction, workload fairness, SLA performance.
Test Approach
Functional validation, risk-based regression, exploratory analysis, solver-behavior comparison, non-deterministic outcome testing.
Biggest Risks
Degraded plan quality, unstable allocation behavior, hidden solver regressions, poor performance at scale.
Outcome
Maintained trust in core scheduling decisions by making quality issues visible early and protecting the release pipeline from high-impact optimization regressions.
02Operations UILive Scheduling Operations Workbench
Business Goal
Drove quality for a real-time operational workspace giving planners, coordinators, and dispatch teams the visibility to control live schedules with speed and confidence.
System / Module
Scheduling views: resources, activities, maps, comparisons, schedule assistance, historical access, live refresh, timezone-aware presentation.
Test Approach
End-to-end UI testing, workflow-based regression, usability-focused validation, cross-view consistency checks.
Biggest Risks
Incorrect operational visibility, broken live refresh, misleading schedule state, inconsistent data across views.
Outcome
Strengthened quality of the main operational workspace used to observe, interpret, and act on live planning decisions.
03Operational ProductivityIntelligent Filtering, Views & Operator Navigation
Business Goal
Supported capabilities helping operational users cut through schedule noise quickly and focus on the resources, activities, timeframes, and constraints that mattered.
System / Module
Filters, bookmarks, saved views, dataset-aware filtering, shared filters, client/server-side filtering, navigation controls.
Test Approach
Functional testing, edge-case validation, data-scope checks, regression around filter combinations, saved states, navigation shortcuts.
Biggest Risks
Hidden data, over-filtered schedules, incorrect scope application, operator error from misleading views.
Outcome
Improved operational efficiency by ensuring users could move through complex schedule data accurately and without unnecessary friction.
04Operational ControlManual Intervention & Schedule Override Controls
Business Goal
Owned quality for the manual intervention layer allowing operations teams to override automation safely when live conditions changed faster than the engine could react.
System / Module
Manual changes, bulk changes, fixed-time and fixed-resource overrides, preview of manual edits, reason capture, change history.
Test Approach
Functional testing, override-priority validation, audit-history verification, regression across high-risk intervention scenarios.
Biggest Risks
Invalid overrides, accidental schedule corruption, non-traceable operator actions, conflicts between manual edits and optimization logic.
Outcome
Made manual intervention safer, more transparent, and more operationally reliable under real-world disruption.
05Decision SupportSchedule Explainability & Decision Diagnostics
Business Goal
Supported an explainability layer to increase stakeholder trust by making automated scheduling decisions easier to understand, challenge, and troubleshoot.
System / Module
Explanation pages, explanation outcomes, diagnostic navigation, linked reasoning flows, schedule-assist behaviors.
Test Approach
Scenario-driven validation, exploratory analysis, message interpretation checks, regression around explanation-linked navigation.
Biggest Risks
Opaque optimization behavior, misleading explanations, weak traceability from symptom to cause, low operator trust.
Outcome
Improved the platform's ability to justify scheduling outcomes and supported faster diagnosis when automated results looked wrong.
06Routing & OptimizationTravel Optimization & Routing Intelligence
Business Goal
Worked on travel-aware optimization to lower non-productive travel time, reduce operational cost, and keep route decisions realistic enough for live execution.
System / Module
Pre-calculated travel matrices, straight-line travel, real-time options, expected arrival logic, travel profiles, route cost behavior, travel analysis.
Test Approach
Route-validation scenarios, cost-model testing, integration-aware regression, performance-sensitive testing on travel-heavy schedules.
Biggest Risks
Inaccurate travel estimates, unrealistic route choices, degraded scheduling quality, performance slowdowns.
Outcome
Ensured the optimizer's routing inputs were credible enough to support reliable field execution decisions.
07Booking PlatformAppointment Booking Engine
Business Goal
Owned quality for a booking engine built to expose viable appointment options quickly, increase customer choice, and translate accepted slots into executable schedule commitments.
System / Module
Appointment requests, offers, offer responses, blocking/non-blocking reservation logic, confirmation flows, booking constraints, concurrency handling.
Test Approach
End-to-end scenario testing, concurrency validation, API/interface checks, negative-path testing for stale or invalid confirmations.
Biggest Risks
Double booking, stale offers, invalid confirmations, downstream failures where accepted slots could not be honored.
Outcome
Improved confidence in customer-facing booking flows by validating both speed and reliability of slot-offer behavior.
08Rules EngineTime Constraints & SLA Enforcement
Business Goal
Supported core timing controls designed to keep work inside valid service windows, contractual promises, and operational availability limits.
System / Module
Scheduling windows, SLAs, availabilities, availability patterns, shift boundaries, committed-activity time rules.
Test Approach
Boundary testing, rule-combination testing, exception-oriented regression, scenario validation across overlapping time controls.
Biggest Risks
Work scheduled outside valid windows, silent SLA breaches, conflicting rule combinations reducing feasibility.
Outcome
Protected service-timing integrity across one of the most business-critical parts of the scheduling model.
09Workforce MatchingResource Preference, Skills & Region Constraints
Business Goal
Worked on the constraint framework ensuring work was assigned to the most appropriate resources based on skill, region, preference, and time-aware eligibility.
System / Module
Resource preferences, region constraints, skill constraints, availability-linked preferences, override priorities, preference-combination behavior.
Test Approach
Rule validation, matrix-style scenario testing, precedence testing, exploratory checks around mixed hard and soft constraints.
Biggest Risks
Invalid eligibility decisions, incorrect override precedence, assignments that violated business intent.
Outcome
Improved trust that the platform was matching work to the right people under real operational constraints.
10Business PrioritizationCost-and-Value Optimization Model
Business Goal
Supported the scoring model that aligned schedule outcomes with business priorities by rewarding high-value work and penalizing expensive execution paths.
System / Module
Activity values, resource costs, travel costs, SLA behavior, activity prioritization mechanics, score-sensitive scheduling outcomes.
Test Approach
Scenario-based scoring validation, sensitivity testing around cost/value settings, regression on priority-driven schedule behavior.
Biggest Risks
Poor business alignment, low-value work crowding out high-value work, unintuitive schedules from weak cost calibration.
Outcome
Ensured optimization decisions reflected business intent rather than accidental configuration.
11Dependency ManagementLinked Activities & Dependency Orchestration
Business Goal
Worked on dependency logic to coordinate related work so multi-step, ordered, or synchronized tasks could execute reliably.
System / Module
Pre-requisites, co-requisites, combined activities, ordered activities, activity pools, separation rules, linked timing.
Test Approach
Chain-based scenario testing, overlap validation, sequence testing, regression across multi-activity orchestration rules.
Biggest Risks
Broken order logic, impossible dependency chains, invalid overlap, fragmented execution of related work.
Outcome
Strengthened confidence in coordinating interdependent operational tasks instead of treating them as isolated jobs.
12Advanced SchedulingSplit Visit Scheduling Framework
Business Goal
Supported scheduling behavior making long-running work operationally deliverable by allowing activities to span multiple visits without losing control or continuity.
System / Module
Splittable activities, minimum visit durations, split costs, split priorities, follow-on behavior, visit completion, multi-resource splits.
Test Approach
Multi-shift scenario testing, interruption/resumption validation, dynamic-update testing, regression on split-routing behavior.
Biggest Risks
Lost follow-on visits, broken continuity across shifts, invalid split behavior under changing conditions.
Outcome
Improved confidence in the platform's handling of long-duration work that could not finish in a single visit.
13Execution StateActivity Status Lifecycle & Dispatch Rules
Business Goal
Worked on the operational state model controlling how activities moved from unallocated to committed, accepted, in-progress, and completed execution.
System / Module
Statuses, status ordering, committed activity behavior, dispatch rules, future-shift handling, shift cut-off, route-validity rules.
Test Approach
Lifecycle-path testing, state-transition validation, rule-override testing, regression around dispatch edge cases.
Biggest Risks
Illegal status changes, invalid dispatch commitments, inconsistent route state, loss of plan-to-field fidelity.
Outcome
Strengthened the control layer that connects optimization output to real-world execution.
14Inventory-Aware SchedulingParts-Constrained Scheduling Engine
Business Goal
Owned quality for parts-aware scheduling logic preventing avoidable service failure by ensuring resources only received work they could complete with available stock.
System / Module
Resource parts, vehicle parts, global/shared parts, stock, depots, pick-up activities, replenishment paths, refill-to-capacity behavior.
Test Approach
Stock-flow scenarios, replenishment testing, depot-route validation, regression around part-category rules and stock updates.
Biggest Risks
Impossible assignments, broken replenishment logic, negative stock, incorrect visibility across shared stock models.
Outcome
Improved trust that scheduled work was actually executable from a parts and replenishment standpoint.
15Capacity ManagementCapacity Rules & External Contractor Buckets
Business Goal
Supported capabilities keeping schedules operationally realistic by limiting work over shifts and modeling outsourced capacity in a controlled way.
System / Module
Shift capacity rules, custom-metric limits, rule overrides, bucket resources, contractor allocations, committed-call behavior.
Test Approach
Rule-combination testing, period-based capacity validation, contractor scenario testing, regression on overridden capacity logic.
Biggest Risks
Over-allocation, invalid capacity exceptions, rule conflicts, unrealistic contractor abstraction dependence.
Outcome
Strengthened schedule realism in environments where workload had to stay within hard operational capacity boundaries.
16Multi-Resource ExecutionParallel Resource & Crew Coordination
Business Goal
Worked on scheduling behavior for work requiring multiple resources to act together, without sacrificing timing accuracy or execution control.
System / Module
Parallel resources, synchronized execution, co-requisite timing, shared end-time scenarios, collaborative route behavior.
Test Approach
Synchronized multi-resource scenarios, edge-case sequencing, regression around timing alignment and status handling.
Biggest Risks
Desynchronized execution, broken multi-resource coordination, invalid joint-visit behavior in live plans.
Outcome
Improved confidence in advanced execution models where a task could not be delivered by a single resource alone.
17Planned MaintenanceCyclic Maintenance Scheduling
Business Goal
Supported recurring-work scheduling to keep repeat maintenance and pattern-based operations consistently planned over time with minimal manual effort.
System / Module
Modelling patterns, repeatable activities, intervals, leeway handling, usage models, operations, cyclic operation scheduling.
Test Approach
Pattern-based testing, recurrence validation, long-horizon scenario testing, regression around usage-driven maintenance logic.
Biggest Risks
Recurrence drift, invalid future timing, incorrect operation sequencing, unreliable repeat scheduling under changing conditions.
Outcome
Ensured the platform could handle maintenance-style work that depended on disciplined repetition rather than one-time dispatching.
18Workforce PlanningWorkforce Planning Foundation
Business Goal
Owned quality across a planning workspace turning workforce structure, planning data, and shift design into a reliable foundation for downstream scheduling.
System / Module
Planning data screens: activity types, resources, skills, SLAs, positions, requirements, parts, rule collections, shift data.
Test Approach
Functional validation, master-data regression, workflow testing, import/export coverage for planning administration.
Biggest Risks
Broken planning master data, invalid setup states, weak data quality feeding downstream execution.
Outcome
Established a dependable planning baseline so scheduling quality was not undermined by bad foundational data.
19Coverage PlanningRole, Position & Requirement Planning
Business Goal
Supported requirement-driven planning to ensure the right mix of positions and roles were available at the right time to meet operational demand.
System / Module
Requirements, role/position requirements, alerting behavior, requirement matching, planning visibility tied to coverage.
Test Approach
Requirement-based scenario testing, under-coverage validation, alert testing, regression across role/position matching logic.
Biggest Risks
False coverage signals, undetected staffing gaps, invalid requirement satisfaction, poor planning from inaccurate data.
Outcome
Improved confidence in workforce coverage planning before shortages could affect live service delivery.
20Team SchedulingTeam-Based Resource Planning
Business Goal
Worked on team-centered planning to let organizations schedule groups of people more effectively where work was delivered as a coordinated team effort.
System / Module
Teams, team membership, memberships over time, integration of teams with scheduling, team-oriented planning behavior.
Test Approach
Lifecycle testing, membership-period validation, inheritance and integration checks, regression around team-based planning flows.
Biggest Risks
Broken membership timing, inconsistent team behavior, invalid downstream assumptions when teams fed live schedules.
Outcome
Strengthened reliability of team-oriented planning where the unit of execution was larger than an individual resource.
21Planning AutomationAutomated Rota Generation
Business Goal
Supported automated rota generation to reduce manual planning effort while producing workable shift structures against real demand and rule constraints.
System / Module
Automated planning, provisional rotas, role overrides, advanced planning options, import shift data, validation on generated plans.
Test Approach
Generated-plan review, rule-conformance validation, regression on planner options, scenario testing around rota feasibility.
Biggest Risks
Invalid generated rotas, weak rule adherence, misleading provisional results, excessive manual cleanup after automation.
Outcome
Improved trust in automated planning outputs by validating that generated rotas were both workable and explainable.
22Governance PlatformArchive, Audit & Recovery Platform
Business Goal
Owned quality for the historical-control layer designed to preserve a full record of schedule changes, enable forensic analysis, and support recovery from major issues.
System / Module
Snapshots, audits, audit browsing, restore flows, timetables, failed audits, archive administration, data import/export.
Test Approach
Data-integrity testing, restore validation, retention-path testing, audit traceability checks, operational recovery scenarios.
Biggest Risks
Missing historical evidence, invalid restore paths, inconsistent archive state, weak post-incident diagnosis support.
Outcome
Strengthened platform governance by ensuring historical state could be trusted, reviewed, and recovered when needed.
23Analytics PlatformHistorical Reporting & Adaptive Analytics
Business Goal
Supported reporting capabilities converting archived scheduling data into KPI-ready, query-friendly insight to feed smarter future scheduling behavior.
System / Module
Reporting tables, fact and dimension structures, reporting generation, analysis data, predictive duration inputs, archive-linked analytics.
Test Approach
Data-pipeline validation, report-structure testing, consistency checks between operational and reporting data, regression on generated outputs.
Biggest Risks
Inaccurate KPIs, broken dimensional analysis, stale reporting outputs, weak trust in history-based improvement signals.
Outcome
Improved confidence in the analytics layer used to understand performance and guide future optimization decisions.
24Simulation PlatformWhat-If Workforce Simulation
Business Goal
Worked on simulation capability letting planners test how changes in demand, resourcing, and targets would affect operational performance before making live changes.
System / Module
Demand modeling, resourcing modeling, automatic resourcing changes, targets, basic/advanced tests, result views, baseline investigation flows.
Test Approach
Scenario-based simulation testing, target-driven validation, baseline-versus-variant comparison, result-interpretation checks.
Biggest Risks
Misleading forecasts, weak scenario realism, false performance assumptions, decisions based on invalid simulated outcomes.
Outcome
Made simulation credible as a decision-support tool rather than just a demonstration feature.
25Scenario ModelingOperational Scenario Visualization & Test Data Simulation
Business Goal
Supported a visual scenario environment letting users create resources, activities, constraints, and test runs quickly before rolling anything into live operations.
System / Module
Scenario creation, bulk data generation, map-driven setup, co-requisites, appointments, imports/exports, data validation, replay-style simulation runs.
Test Approach
Workflow testing, scenario-integrity checks, simulation-run validation, regression across visual and imported test data paths.
Biggest Risks
Invalid generated data, misleading visual simulations, weak scenario portability, poor fidelity between simulated and operational behavior.
Outcome
Improved confidence in the platform's ability to model, test, and demonstrate complex scheduling behavior safely outside live production.
Root Cause Analysis

Problem-Solving Case Studies

Complex defect investigations and high-impact resolutions that prevented critical production failures.

01
Non-Deterministic AI Scheduling Output Producing Invalid Plans
Problem

The AI PSO scheduling engine was producing plans that passed basic functional checks but violated real-world business constraints — over-allocated resources, infeasible task sequencing, and cost estimates outside acceptable bounds. Standard test cases couldn't catch it because no two outputs were identical.

Investigation

Existing test cases validated specific expected outputs — inapplicable for non-deterministic engines. Mapped the domain's hard constraints (resource capacity, task dependencies, cost ceilings) and designed invariant-based checks that held true regardless of the specific schedule produced.

Resolution

Pioneered a constraint-driven validation framework — rules, invariants, feasibility checks, cost bounds, and impact assessments — that validated outputs against business rules rather than exact values. Adopted as the team's standard methodology for all AI output testing.

Zero AI-related production escapes across 90+ sprints. Framework became the R&D standard for non-deterministic validation.

02
Critical Data-Layer Inconsistencies Hidden Beneath Passing UI Tests
Problem

Functional UI tests on the IFS platform were consistently green, yet stakeholder-reported data discrepancies in scheduling outputs persisted across releases. Root cause was unknown and intermittent.

Investigation

Shifted focus below the UI layer. Used Postman for systematic API contract validation and deep SQL queries to verify data integrity across service boundaries. Identified a silent data transformation mismatch at an integration point that corrupted values before they surfaced in the UI.

Resolution

Introduced mandatory API + SQL validation layer as a standard gate in the release checklist. All cross-service data flows now verified at the data layer, not just through UI assertions.

Eliminated the class of silent data bugs. End-to-end data correctness verified across all services and integrations from that sprint forward.

03
Uncontrolled Defect Ageing Blocking Release Pipeline
Problem

Defects were accumulating across sprints with inconsistent severity classifications, unclear ownership, and no enforced resolution timelines. The release pipeline was being blocked by ambiguous defect states and stakeholders lacked confidence in quality reporting.

Investigation

Audited the full open defect backlog. Found that 60%+ of open items lacked reproducible steps, clear severity justification, or assigned resolution accountability — making triage decisions arbitrary and slow.

Resolution

Led 80+ defect triage and quality gate review sessions, establishing reproducibility standards, severity discipline, and clear Dev accountability. Built Jira & Xray dashboards tracking defect ageing, trends, and coverage — producing weekly stakeholder-ready release reports.

Average defect age cut by 35%. Go/no-go decisions became data-driven. Release pipeline unblocked and on-time delivery sustained across 90+ sprints.

QA Metrics

Interactive Impact Dashboard

Real numbers from 3+ years of quality ownership at IFS R&D.

0
Agile Sprints Delivered
3+ years · sole QA lead · consistent on-time delivery
0
Defects Governed to Closure
Triaged, classified & resolved · avg age cut 35%
0
Test Cases Executed
510+ user stories · full spec-to-outcome traceability
0
Tech Specs Reviewed
Shift-Left: flagged edge cases before development began
40% prevented
Defects Prevented
via Spec Reviews
35% reduction
Avg Defect Age
Cut
25% less rework
Review Rework
Reduced (Mentoring)
ZERO escapes
Production Escapes
90+ Sprints · 0 Defects
Zero Production Escapes

Sole QA lead across 90+ sprints on a non-deterministic ERP & AI scheduling platform — achieved full traceability from specs to outcomes with zero defects escaping to production across the entire engagement.

0
Agile Sprints
Delivered

3+ years as sole QA lead on IFS ERP & AI PSO scheduling platform. Consistently on-time, high-confidence releases across every sprint.

0
Defects Governed
to Closure

Triaged, classified, and accelerated resolution for 1,200+ defects — enforcing reproducibility standards and severity discipline across Dev, Product, and stakeholders.

0
Test Cases
Executed

Engineered and executed 2,500+ test cases across 510+ user stories spanning functional, regression, integration, smoke, and UAT — with full spec-to-outcome traceability.

0
Tech Specs
Reviewed

Reviewed 67+ technical specifications and design changes before development, flagging edge cases, testability gaps, and integration risks — preventing an estimated 40% of defects from ever being written.

40% prevented
Defects Prevented via Spec Reviews
Shift-Left technique — flagging issues at design stage before a single line of code was written.
35% faster
Reduction in Avg Defect Age
Structured triage, reproducibility standards, and severity discipline cut the time defects stayed open.
ZERO escapes to prod
Production Escapes
Zero critical defects reached production across 90+ sprints.
Quality Gate Discipline

80+ defect triage and quality gate reviews conducted as sole QA lead — aligning Dev, Product, and stakeholders on every release decision with data-driven go/no-go reporting.

Defects Prevented via Shift-Left Spec Reviews 0%

Reviewing 67+ tech specs before development flagged edge cases, testability gaps, and integration risks early — preventing an estimated 40% of defects from ever entering the codebase.

Reduction in Average Defect Age 0%

Structured triage governance — enforcing reproducibility standards, severity discipline, and clear Dev ownership — cut the average time defects remained open by 35%.

Review Rework Reduced via QA Mentorship 0%

Mentoring and upskilling 3 senior QA engineers raised test design quality and defect-writing rigour, reducing the volume of work sent back for correction by 25%.

Release Triage Meetings Led as Sole QA 0

Led 80+ defect triage and quality gate sessions — producing stakeholder-ready Jira & Xray dashboards that directly informed go/no-go release decisions for enterprise products.

Leadership & Validation

Beyond the Test Cases

Building quality cultures, leading teams, and earning industry recognition.

QA Mentorship

Mentored and upskilled 3 senior QA engineers at IFS, raising test design quality, defect-writing rigour, and execution discipline — reducing review rework by 25%.

Stakeholder Management

Produced stakeholder-ready release reports (coverage, defect trends, ageing, execution status) that directly informed go/no-go decisions for enterprise product releases at IFS.

Quality Gate Ownership

Led 80+ defect triage and quality gate review sessions as sole QA lead — aligning Dev, Product, and stakeholders on release priorities and systematically unblocking pipelines.

AI-Assisted QA Innovation

Championed AI-assisted QA adoption (test design acceleration, regression optimisation, defect clustering, root-cause support) and partnered with Architecture & R&D to modernise QA ways of working across the organisation.

ISTQB CTAL-TM v3.0 — 2025
Lean Six Sigma Green Belt — 2019
Certificate of Software Automation — 2023
Advance Diploma in JAVA — Uni of Kelaniya
MBA in Entrepreneurship — Anglia Ruskin University
M.Sc. Computer Science — Uni of Sri Jayewardenepura
B.Sc. Electronics & Automation — Uni of Colombo

Awards & Community

★★★★★

Roshan consistently delivered on-time, high-confidence releases across 90+ agile sprints as our sole QA lead on the IFS PSO platform — a non-deterministic AI scheduling engine where most engineers wouldn't know where to begin with testing. His constraint-driven validation framework became our R&D standard.

IFS R&D Engineering Team
IFS R&D International Pvt Ltd
★★★★★

Roshan's Shift-Left approach transformed how our team thinks about quality. By reviewing specs before development even begins, he prevented an estimated 40% of defects from ever being written — saving us sprint cycles and release risk we didn't even know we were carrying.

Product & Dev Leadership
IFS R&D International Pvt Ltd
★★★★★

What makes Roshan exceptional is his ownership mentality. He doesn't just execute test cases — he architects quality strategy, mentors the team, manages stakeholder expectations, and drives process improvements all at once. A rare combination of technical depth and communication skill.

Senior Engineering Stakeholder
IFS R&D International Pvt Ltd