Christopher Lam

Worked on building a TypeScript web platform for end-to-end project management for solar panel retailers, including tracking of solar tax credits. Built and orchestrated AI agentic systems using Claude, deployed on AWS/Google Cloud to run headlessly with browser-style automation, including programmatic web login flows and form submission/download of external assets. Delivered STC agent workspace features including evidence, compliance, RFI and declaration views, and fine-grained permissions via an invite system. Created backend services for live chat streaming and long-running tasks using SSE, plus APIs to expose progress, active agents, tool calls, assumptions, errors, and completion state. Integrated Xero settlement with feature flags, batched pushes, rate limits, and live sync status; implemented SalesForce and HubSpot integrations via OAuth. Implemented handoff patterns, evaluations, guardrails, and agent toolsets using OpenAI Agents SDK.

Developed Python pipelines on HPC GPU clusters to queue protein prediction jobs using AlphaFold and CombFold. Conducted combinatorial assembly by dividing complex proteins into dimer chains for FASTA inputs using NumPy and pandas. Created PAE, sequence coverage, and pLDDT position visualizations in Jupyter to compare and interpret predicted vs experimental protein structures.

Cleaned and transformed large CSV datasets containing clonal cell count data. Applied machine learning techniques such as PCA and MDS for feature extraction and generated visual plots for RNA-seq data. Automated bar plot visualizations using seaborn and filtered data for export to RStudio.