Daniel Fridljand

How I Used Claude Design to Create a Hackathon Slide Deck

A short post about using Claude Design to create a hackathon slide deck, with the full deck embedded below.

Architecture of a First-Level Support Automation

The four-step pipeline we landed on after iterating in production: a rooted DAG of intents, grouped extraction, weighted-scoring fetch, and a mostly-hardcoded planner with an explicit LLM fallback.

On the Future of Pull Requests in the Age of Agentic Coding

Masks and Cross-Attention: Extending Self-Attention

Encoder, Decoder, Both: The Three Transformer Architectures

Self-attention is the mechanism. BERT, GPT, and T5 are three different ways to stack it.

From Kernel Regression to Self-Attention: A Line-by-Line Derivation

Agentic vs Workflow-based AI

Workflow win over agents in business process automation.

Book Review: Agentic Design Patterns: A Hands-On Guide to Building Intelligent Systems

Practical, critical review of 'Agentic Design Patterns' — evaluates agent design techniques, code examples, and their real-world usefulness for engineers and researchers.

Unite x TUM.ai Data Mining Hackathon: Core Demand Prediction

How we tackled Challenge 2 at the Unite x TUM.ai Data Mining Hackathon: fee-aware portfolio optimization for recurring procurement demand, from raw PLI data to a two-stage model and NACE-informed cold start.

Langfuse in Production: Monitoring, Persistent Traces, and a Self-Improving Codebase

How I use Langfuse in an AI support pipeline for monitoring, persisting evaluation runs for later review, closing the loop with a multi-agent triage loop driven by Cursor, and a decision framework for when to fix eval data versus prompts.

Type-Safe Hybrid Workflows with Pydantic AI

One Pydantic AI primitive scales across deterministic, LLM-assisted, and agentic steps — with the same typed contract everywhere.

Temporal for Human-in-the-Loop: When You Don't Know How Long to Wait

How Temporal's durable waits, signals, and deterministic recovery make human-in-the-loop workflows reliable when approval times are unknown.

Strategic Model Selection in Cursor: Balancing Cost and Performance

Learn how to optimize your Cursor usage by choosing expensive models for planning and cheap models for execution, with a cost vs performance chart updated daily.

Orchestrating ML Data Pipelines with Snakemake: A Real-World Case Study

How I used Snakemake to manage a complex 19-stage ticket analysis pipeline, leveraging scatter-gather parallelization to efficiently process 2000+ customer support tickets with LLM classification.

Prompt Injection

I participated in a white hat hacking challenge at Agent Olympics Hackathon Munich 2026. Can you trick my chatbot into revealing the secret password?

Stop Prompting, Start Programming with DSPy

DSPy is a framework that separates the 'what' from the 'how' in LLM development, moving from brittle prompt engineering to robust software engineering.

Useful Resources for AI Agents

A curated collection of essential resources for building and deploying AI agents, covering frameworks, best practices, and engineering insights from leading organizations.

Building My Personal Website: Technical Highlights

A deep dive into the technical implementation of my personal website, covering the AI chat interface, markdown context generation, and resume PDF sync from a dedicated CV repo.

Extending HDP for Mutational Signatures: Incorporating Evolutionary Trees

An extension of the Hierarchical Dirichlet Process model that incorporates phylogenetic tree structures to capture evolutionary relationships in mutational signature activities.

Mutational signature estimation with Hierarchical Dirichlet Process

A mathematically rigorous introduction to mutational signature estimation using Hierarchical Dirichlet Processes, bridging the gap between biological applications and statistical methodology.

Deep Learning-Based Diabetic Retinopathy Detection in Fundus Images

A deep learning approach for diabetic retinopathy detection using fundus images, combining binary classification and image segmentation to identify disease presence and quantify key indicators.