Mushaim Khan

Designed and tuned the AI chatbots conversational behavior — defined its response preferences,
tone, and guardrails (what it should and shouldnt say, and how it should answer) through prompt
engineering and iterative testing — improving response consistency and keeping replies on-topic and accurate (~90% on-intent responses).
Built multi-agent LLM workflows (role-separated agents: drafting → peer-review → reporting) that hand off structured output automatically, improving reliability and removing manual hand-offs.
Engineered Python automation and scheduled data pipelines integrating multiple REST APIs into a single production flow, with auth, pagination, and retry/error handling for reliable unattended runs (near 100% successful automated runs).
Owned the automation stack end to end — ingestion, processing/flagging logic, and delivery — cutting repetitive manual review for the team by an estimated ~70-80% on automated tasks.

Collected, cleaned, and structured large datasets from multiple REST-API and operational sources, ensuring quality and readiness for analysis.
Built and validated anomaly/fraud-detection models with explainable flagging, and delivered
automated dashboards and weekly reports — cutting manual review by ~6-10 hours/week per client.
Ran exploratory data analysis to surface trends and risks, and presented findings to non-technical stakeholders in clear, actionable terms.
Designed QA / model-evaluation frameworks (consistency benchmarking, edge-case coverage) to guarantee data quality before delivery.

Managed an affiliate program end to end — onboarding partners, setting up and troubleshooting
conversion tracking (UTM attribution and tracking pixels), and verifying charges/commissions for accuracy.
Automated the entire process: built a system that ingests performance data, attributes conversions, flags fraudulent/suspicious partner activity with explanations, and auto-generates reports — cutting ~6 hours/week of manual review and improving accuracy.
Fixed and adjusted HTML/CSS in creative templates, coordinated across teams, and maintained clean, detail-checked records.

Built Python automation and scheduled data pipelines integrating REST APIs (ingestion → processing → reporting) that run reliably unattended.
Delivered full-stack web work (HTML/CSS/JavaScript, React/Next.js) and AI integrations across multiple client projects.

Built scheduled Python pipelines that ingest live REST-API data, apply segmentation/anomaly logic, and auto-generate reports — cutting manual operational review by ~10 hours/week per client.
Engineered an affiliate-tracking + fraud-detection system with explainable flagging and fully automated monthly reporting — saving ~6 hours/week of manual review and surfacing risky activity that was previously missed.
Built full-stack web applications and AI integrations end to end (frontend, REST APIs, deployment) across multiple client projects. QA / regression frameworks LLM-driven pipelines output-consistency benchmarking edge-case