SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

How to Run Separate Personal and Work Claude Code Accounts on One Machine

0
·4 views

A developer has shared a simple method for isolating two Claude Code accounts on the same machine, keeping personal and work environments completely separate. The approach relies on Claude Code's built-in support for the CLAUDE_CONFIG_DIR environment variable, which redirects the app to a different config directory. By defining two shell functions in .zshrc, each pointing to its own config folder, the user ensures credentials, MCP servers, hooks, and session history never mix between accounts. The developer notes this avoids risks like work tokens leaking into personal sessions or internal MCP server hooks firing in the wrong context. While the trade-off is that settings must be updated in both accounts independently, the author considers this a minor inconvenience compared to the security and clarity gained.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Engineering team cuts RAG pipeline latency 40% using Bayesian search and hybrid retrieval

A development team rebuilt their retrieval-augmented generation (RAG) pipeline from the ground up after standard production deployments proved inadequate for legal, API, and support-ticket use cases. The team replaced fixed 512-token chunking with document-specific strategies — including recursive, semantic, and agentic chunking — achieving recall@10 scores between 91% and 97% across content types. They combined vector search with BM25 and cross-encoder reranking using Reciprocal Rank Fusion, which improved relevance correlation from roughly 0.75 to 0.92. A Bayesian hyperparameter search was used to tune retrieval weights, ultimately cutting end-to-end query latency by 40%. The approach demonstrates that measurable, tunable retrieval pipelines significantly outperform default RAG configurations in real-world production environments.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How One Team Built Automated LLM Evaluation That Catches 92% of Hallucinations

A software team discovered critical gaps in their AI quality process after deploying a RAG-based customer support assistant that served hallucinated responses to over 500 users before detection. The failures included fabricated billing policies and rate-limit figures pulled from a competitor's documentation, exposing a test process that relied entirely on manual spot-checks. In response, the team built a production-grade evaluation pipeline integrating automated judges for faithfulness, instruction-following, JSON schema validation, and domain accuracy into their CI/CD workflow. The system uses an ensemble of LLM-based and deterministic judges with defined score thresholds to block code merges that degrade response quality. The approach replaces informal human review with versioned golden datasets and regression detection, enabling systematic quality control before any model changes reach production.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Compressing an Image to an Exact File Size Is a Hard Engineering Problem

JPEG compression does not directly control file size — it adjusts a quantization table that discards image detail, meaning the same quality setting can produce vastly different file sizes depending on image content. A smooth sky may compress to 15KB at quality 70, while a detailed crowd photo at the same setting could exceed 180KB. The file-size-to-quality relationship is also non-linear, flattening at lower quality levels where further reductions yield diminishing returns. Binary searching the quality parameter is a common workaround but fails when images cannot reach a target size without also reducing resolution or when very low quality produces unusable artifacts. A robust solution must tune multiple parameters together — including dimensions and chroma subsampling — rather than relying on quality alone.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer anxiety grows as AI coding agents threaten to automate routine engineering tasks

A software developer has voiced growing career anxiety over the rapid advancement of AI coding agents. Tasks such as boilerplate CRUD operations, basic API endpoints, and UI scaffolding — which once consumed significant sprint time — can now be generated in seconds. The developer fears that routine daily work could be fully automated within two years. This concern is prompting reflection on how engineers can build differentiated, harder-to-automate skills to remain relevant long-term.

0 comments Read more at DEV Community