My Work
A mix of things I’ve built, written, and experimented with. ✏️
Automated Quote Generation for a Biotech Firm
An AI tool to analyze bias and sentiment in news articles
Tailor applications to job descriptions in seconds
Extracting insights and contacts from medical research papers
Drove 30% YoY traffic growth through data-informed content
Workflow Automation
Design and implementation of end-to-end automation systems that connect lead acquisition, data processing, business tools, and AI components to reduce manual work and improve operational flow.
- Creation of conversion-focused landing pages.
- Automated capture and validation of inbound leads.
- Custom retrieval and scraping tools for prospect collection.
- Lead enrichment and filtering based on relevance.
- Automated ingestion of leads into CRMs via APIs/webhooks.
- Event-driven notifications for outreach actions.
- Personalized outbound and follow-up email workflows.
- Automation for creating/updating contacts, deals, and pipelines.
- Gmail-based outreach with automatic logging.
- Webhook-driven notifications for follow-ups and reminders.
- Data validation, deduplication, and normalization for clean CRM records.
- Automated data aggregation from multiple sources.
- Scheduled generation of reports and summaries.
- Google Sheets automation for formatted reports.
- Alerts to stakeholders when reports are ready.
- Custom AI-powered chatbot assistants using Gemini agents.
- Context-aware responses grounded in internal data.
- Tool-enabled assistants for actions like data lookup or CRM updates.
- Retrieval-based workflows to improve reliability.
- Escalation logic for edge cases and human handoff.
Event-Driven
Architecture using webhooks for real-time responsiveness.
Layered & Separated
Clear separation between data ingestion, processing, and action layers.
Reliable & Transparent
Focus on reliability and low-friction user experiences.
Extensible
Designed for extensibility and real-world operational use.
"To accompany the growth of our company, we have to reinforce our prospect efforts toward targeted customer segments, particularly academic researchers. In this approach, we had to generate a very specific database of published authors, matching our fields of application. To this end, Yajush developed a parser to be used on the largest published article database (PubMed), allowing us to generate an exploitable file (CSV) containing information on the authors' contacts and related articles. The parser is made in a way that it is easily deployable, based on API, and allows filtering on various parameters. These combinations of features are essential to customise as much as possible the emails to be sent, and to be very specific in the targeted prospects."