venessa.nalugala@strathmore.edu
24 yrs oldFull-Stack Software Engineer with 5+ years of experience building production-grade systems across Python
backends, React/Next.js frontends, and AI-powered applications. Deep expertise in FastAPI, Django REST
Framework, TypeScript, PostgreSQL, and cloud infrastructure on AWS and GCP. Proven track record integrating
LLMs, RAG architectures, and AI agents into production systems. Experienced in async programming, SQL
optimization, and shipping scalable, maintainable software in fast-moving startup environments.
Bachelor's Degree
5+ Years (Senior)
DANI (Document Analysis & Natural Language Interface)
DANI is an AI-powered executive assistant web app built for VGG and now being expanded to other clients, starting with RedTech. It replaced a team of 6–7 executive assistants who listened to every meeting recording by hand and typed action items into Notion. DANI connects to Fireflies.ai through webhooks and a 30-minute background sync, so new meeting transcripts are pulled in automatically. The AI then extracts structured action items, each with an owner, deadline, priority, confidence score, and the exact quote from the meeting as proof. Validation rules reject vague or duplicate tasks, and items the AI is unsure about go to a human review queue. Users can also chat with DANI to ask questions about their meetings and documents, such as “What did John promise about the budget last Tuesday?” Answers come with citations to the exact meeting and timestamp, using hybrid semantic and keyword search. Over time I delivered additional modules: a deal CRM pipeline, contacts intelligence, a WhatsApp bot for voice updates, a Ghostwriter tool that drafts content in the user’s own style, and multi-tenant support for onboarding new organisations. The stack is Next.js/React, FastAPI, PostgreSQL, Qdrant, Redis, and AWS S3, deployed on Render and AWS.
Portfolio Indexing System (PI)
PI is an AI-driven knowledge base for an investment firm. It brings all portfolio company data (pitch decks, financials, and web and LinkedIn intel) into one searchable hub. Before PI, a simple question like “Which of our companies are hiring in East Africa?” took days of digging through files. Now the investment team asks in plain English and gets cited answers in seconds, across one company or the whole portfolio. The app has three portals. Portfolio companies onboard themselves and keep their own profiles updated. Analysts get portfolio-wide AI search with role-based access. Admins get audit trails, partner invitations, and cost “kill switches” for the enrichment automations. Key features include a document ingestion pipeline that extracts, chunks, and embeds uploads for retrieval-based search. Agentic enrichment using Gemini and LinkedIn scraping fills profiles automatically, and a rules engine scores each profile for completeness and generates targeted forms to fill the gaps. The system also includes subscription billing, GDPR-compliant encryption, and an admin override feature to correct inaccurate scraped data. I took it through real portfolio data ingestion and a live demo to MTN. It’s built on React, FastAPI, PostgreSQL, ChromaDB, AWS S3, Cognito, and EventBridge.