Job Search Status
Open to workDesired start date: Flexible
Open To Work
Tekla Khojava
Tbilisi
Work Preference
Summary
Detail-oriented AI/LLM QA Specialist with a focus on enhancing the reliability of cloud-based conversational platforms through rigorous testing and monitoring. Identifies edge cases and enforces quality thresholds with evidence-backed triage. Collaborates with AI product and data teams, leveraging expertise in REST/JSON and CRM/Salesforce validation while advancing skills in Python automation and Azure.
Overview
3
3
years of professional experience
2
2
Certifications
Work History
Software Testing Engineer
EPAM Systems
10.2025 - Current
- Delivered hands-on exercises and best-practice playbooks for AI-assisted exploratory testing, prompt behavior checks, and evidence capture.
- Delivered practical exercises and hands-on support to 50 QA professionals, facilitating adoption of AI-assisted testing practices.
- Advised teams on AI adoption strategies and QA use cases, raising testing efficiency and consistency across projects.
- Evaluated responses of an educational AI platform, designing and writing prompts, submitting detailed reports, and logging precise bug reports with clear, reproducible steps.
- Assessed AI outputs for accuracy, relevance, and alignment with learning objectives, providing structured feedback to improve system performance.
- Upskilling in Python, test automation, and AI integrations through EPAM internal programs.
QA Specialist (Functioning as AI QA Engineer)
Impel AI
02.2023 - 10.2025
- Owned end-to-end QA for production conversational AI (Sales AI, Chat AI, Voice AI, Knowledge Bank) supporting 8,000+ dealerships across 53 countries.
Designed and executed test cases for non-deterministic LLM behavior using acceptance criteria, heuristics, and quality thresholds; validated outputs against product rules and safety guidelines.
- Owned intent-level testing, prompt behavior validation, fallback logic, and AI response quality, ensuring tone, accuracy, and minimizing hallucination risks.
- Built and maintained regression suites for AI features (intent/prompt/workflow), cutting critical AI issues by ~80% through edge-case discovery (hallucinations, inconsistency, context loss) and targeted fixes.
- Conducted deep API/payload analysis (REST, JSON, Postman) to identify root causes in intent extraction, data mapping, and logic flows; documented reproducible steps, logs, and evaluation evidence.
- Validated conversation flows, voicemail grouping, default values, and CRM/Salesforce integrations to ensure cross-system consistency and reliability.
- Performed post-launch quality analysis on live conversations, achieving an approximate 80% reduction in critical AI issues over time.
- Monitored post-launch AI KPIs and system behavior, recommending optimizations for sustained quality.
- Served as the primary QA liaison across Product, Engineering, Logic, Conversation Design, and Data Operations teams.
- Recognized as a subject-matter expert for complex AI behavior during major releases, migrations, and high-risk launches.
- Authored clear, actionable test documentation informing prompt/logic updates and risk mitigation for high-risk releases and migrations.Led end-to-end QA for production conversational AI systems (Sales AI, Chat AI, Voice AI, Knowledge Bank), supporting 8,000+ dealerships across 53 countries.
- Ensured seamless integration between AI systems and dealership CRMs/Salesforce.
- Tested CMS implementations for chat AI dashboards and evaluated AI-generated responses for accuracy and relevance.
- Contributed to AI test strategy development and exploratory testing practices for internal projects.
Skills
AI testing and quality assurance
- API testing with REST protocols
- JSON payload accuracy analysis
- Quality thresholds and acceptance criteria
- Workflow regression testing strategies
- Edge case discovery in AI outputs
- Safety & guardrails: prompt guardrails, safety policy compliance, red teaming
- Data quality & drift: drift detection, bias/fairness checks, dataset curation/labeling
- Non-deterministic system evaluation
- Heuristic and semantic validation
- Monitoring AI response metrics
- Intent and fallback logic validation
- Business rules compliance behavior
- Post-launch degradation alerts management
- Integration checks across CRM/Salesforce systems
- Evidence capture during issue reproduction
- Clear test plan development with insights
- Cross-functional collaboration for documentation
- Growth Areas: Python and test automation (PyTest – learning), Azure (upskilling), evaluation/observability tools (ready to adopt LangFuse/Ragas/TruLens and telemetry dashboards)
Awards
- Performance Rating, 4.4 - Exceeds Expectations for analytical depth and measurable quality impact - Impel AI, 2024
- Trusted by the Director of Engineering, to independently validate high-risk AI releases
- Improved AI reliability and response quality, at enterprise scale (8,000+ dealerships across 53 countries)
Personal Information
Title: QA Engineer (AI & Integration Systems)
Languages
English
Advanced (C1)
C1
Georgian
Native
Native
Certification
Python for Everybody, University of Michigan / Coursera
Timeline
Software Testing Engineer
EPAM Systems
10.2025 - Current
QA Specialist (Functioning as AI QA Engineer)
Impel AI
02.2023 - 10.2025