P AKASH RAO
Data Analyst
Hyderabad
I’m a Fraud & Risk professional with 5+ years of experience across fraud detection, abuse prevention, risk investigations, data analysis, quality assurance, and process improvement. Currently, I work at Dun & Bradstreet, where I focus on business risk and fraud investigations, identifying suspicious patterns, analyzing risk signals, and improving investigation processes. I have experience working with IBM Trusteer to optimize risk policies, reduce false positives, and improve the identification and tracking of emerging fraud patterns, including indicators associated with shell and shelf businesses. Previously, I worked at Amazon in Abuse Prevention as a Senior Investigator Specialist, where I handled complex investigations involving account abuse, seller abuse, payment-related risk, identity risk, suspicious activity, and policy violations. I also served as a Quality POC and SME, supporting investigation quality, process improvements, and collaboration with stakeholders across multiple regions. Throughout my career, I have worked with large datasets and used analytical approaches to identify trends, anomalies, fraud patterns, and potential risks. I have also focused on automating repetitive investigation workflows to improve efficiency and reduce manual effort. Core Skills • Fraud Detection & Investigation • Risk Analysis & Risk Assessment • Fraud & Abuse Prevention • Trust & Safety • Transaction & Behavioral Analysis • Suspicious Activity Analysis • Root Cause Analysis • Risk Policy & Rule Optimization • False Positive Analysis • SQL & Data Analysis • Python • Advanced Excel • Power BI • AI / Generative AI & Prompt Engineering • Process Automation • Quality Assurance & SOP Compliance • Stakeholder Management I’m particularly interested in solving complex fraud and risk problems by combining investigation expertise, data analytics, AI/GenAI, and automation. I enjoy identifying opportunities where technology and data can improve detection, investigation quality, and
Analyst 2
Dun & Bradstreet
- Conduct investigations on businesses across global markets to identify potential fraud, business misrepresentation,
- business impersonation, and business compromise using internal research and investigative tools.
- Perform detailed business research and risk assessment by analyzing company information, registration details,
- operational characteristics, ownership-related information, and other available signals to identify suspicious or
- inconsistent business activity.
- Conduct open-source research (OSINT) to validate business information, investigate potential risk indicators, and
- support evidence-based case decisions.
- Identify emerging fraud patterns, suspicious business characteristics, and risk trends, escalating relevant findings in
- accordance with investigative procedures and internal standards.
- Research and validate business information across relevant Secretary of State and government business-registration
- sources, supporting verification of business legitimacy and operational status.
- Review high-risk cases and independently perform investigations while maintaining accuracy, quality, documentation
- standards, and case turnaround expectations.
- Analyze multiple data points and information sources to distinguish between legitimate business activity and potential
- business identity theft or material misrepresentation.
- Document investigation findings clearly and provide concise research summaries and risk-based recommendations for
- internal stakeholders and clients.
- Collaborate with cross-functional teams and stakeholders to communicate investigation findings, identify new risk
- patterns, and support initiatives designed to protect the D&B database from fraudulent or materially misrepresented
- business information.
- Leverage AI-driven tools and automation to streamline repetitive research activities, improve investigation efficiency,
- and support identification of emerging fraud patterns.
- Use Microsoft Excel, Power BI, SQL, and Python to support data analysis, investigation workflows, trend
- identification, and process improvement activities.
- Contribute to process-improvement and automation initiatives, identifying opportunities to reduce manual research
- effort and improve operational efficiency within Fraud Operations.
Sr Risk Analyst
Amazon
- Conducted end-to-end fraud and abuse investigations across high-volume workflows, analyzing behavioral,
- transactional, account, and payment signals to identify suspicious activity and emerging abuse patterns.
- Performed fraud pattern and trend analysis across large datasets to identify recurring risk vectors, investigate root
- causes, and support preventive controls.
- Analyzed 10,000+ records monthly using SQL and Advanced Excel, identifying anomalies, relationships, and risk
- indicators to support investigation and enforcement decisions.
- Investigated compromised payment methods, account abuse, seller abuse, and linked-account activity, identifying red
- flags and initiating risk-based enforcement actions.
- Identified emerging abuse patterns and escalated high-risk findings to stakeholders, supporting improvements to
- investigation workflows and detection controls.
- Maintained 98% investigation quality while consistently meeting operational SLAs and documentation standards.
- Applied structured quality analysis and SOP reviews to improve investigation consistency and support scalable
- enforcement workflows.
- Collaborated with cross-functional stakeholders across global operations to communicate investigation findings,
- emerging risk patterns, and process-improvement opportunities.
APEX (ASIN Processing and Execution) Tool
Accomplished a 99%+ reduction in ASIN processing time, reducing manual processing from 5 minutes per ticket to ~2 seconds, by developing an AI-powered PartyRock tool that converted structured inputs into executable Brazil CLI code. Increased processing capacity from ~150 tickets/day to ~4,280 tickets/day, measured through automated execution and reduced manual intervention. Reduced an 8-hour manual workload to ~30 minutes of supervision and enabled a one-week backlog to be cleared in one day, by automating repetitive processing steps. Built and validated the prototype in 1 week, followed by 2 weeks of testing, and launched in the 4th week.
POA Reader – LLM-Based Document Analysis & Abuse Detection
Automated multi-document seller appeal analysis, reducing repetitive manual review by using Python, PartyRock, and LLM-based comparison of root cause, corrective action, and preventive measures. Improved consistency and reduced false positives by designing structured prompts to identify duplicate and near- duplicate Plans of Action. Enabled faster identification of repetitive or potentially abusive appeals, supporting fraud-prevention investigations.
Attendee · Oct 24, 2026