Deniz Cansin Rodoplu

External & Customer Facing Agents Lead Backend Engineer - Lead engineer for JobTarget's customer facing AI agents: Designed multi-agent systems on LangGraph and Deep Agents spanning orchestration, intent classification, ReAct-style reasoning, tool calling over MCP, and agent-to-agent (A2A) communication. - Developed a low-latency, high-throughput serverless runtime on AWS (Lambda, DynamoDB) that scales to zero between sessions — session-scoped short-term memory with cross-session retrieval for long-term recall. - Implemented the security and context-engineering layer end to end: input/output filtering, scope enforcement, prompt harness design, and middleware hardening against jailbreaks, prompt/code/SQL injection, and PII leakage — the controls that made customer-facing deployment viable. Search & Matching - Data Ingestion - Co-built the backfill that enriched and loaded 7M candidate resume records into Vespa, the datastore behind JobTarget's Candidate Manager product. - Own the realtime ingestion path — a distributed, event-driven SNS → SQS → Lambda pipeline enriching and upserting 10K candidate records/day into Vespa, built for horizontal scale with retry and dead-letter handling for guaranteed delivery.

Product lead for Lighthouse Screening, a SaaS case management and screening tool, hosted on AWS, used by 160+ organizations and 3,000+ users across Texas and Louisiana. - Orchestrated end-to-end product development and user adoption. - Acted as a UI/UX designer and crafted UI/UX components on Canva & Figma. - For customer success, implemented comprehensive training and user engagement programs. - Lead key stakeholder meeting with partners to decide on new product features, enhance user engagement, and increase screening. As Data Lead, managed 5 out of 15 key dashboards within Lighthouse Data, utilizing Python, AWS, and SQL to ensure data accuracy, freshness, and consistency. - Oversaw the development of visualizations that effectively conveyed complex scenarios and insights, driving informed decision-making across the organization and external stakeholders. - Implemented the ingestion, processing, integration, and automated cleaning of 50M+ data records across multiple datasets using AWS Aurora (SQL), developing workflows for transformation and validation that reduced manual intervention by 50%. - Using Python and R, performed advanced statistical analysis (including cluster methods, p-tests, A/B tests, and psychometric analysis) to produce in-depth analytical reports (50+ pages), translating complex data insights into actionable recommendations for non-technical audiences.