Build a systematic workflow for evaluating STAR response quality, detecting prompt regressions, and iterating on prompts using data — not gut feel.
Manish Sinha manish
manish
/ langfuse-evaluation-plan.md
Last active
April 7, 2026 02:13
StarPrepare — Langfuse Evaluation & Prompt Iteration Plan
manish
/ ios-code-review.md
Created
February 18, 2026 17:36
manish
/ starprepare-ui-ux-analysis.md
Last active
February 16, 2026 22:41
StarPrepare UI/UX Analysis: Comprehensive Comparison with ChatGPT & Claude (Feb 2026)
StarPrepare UI/UX Analysis: Comprehensive Comparison with ChatGPT & Claude (Feb 2026)
Report Date: February 16, 2026 Analysis Method: Puppeteer screenshots, code inspection, competitive research, expert AI validation (GPT-5.1-Codex) Scope: Web interface (starprepare-93411.web.app), desktop only (1920x1080+)
manish
/ agent-swarm-plan.md
Created
February 1, 2026 20:23
Fire Bikini Body - Specialized Agent Swarm Architecture Plan
A coordinated swarm of 10 specialized agents designed to tackle Fire Bikini Body's open issues, implement missing features, and maintain code quality. Each agent has specific MCP server access tailored to its responsibilities.
This plan covers 25 high-value test cases that humans should verify manually. These focus on UX quality, visual polish, and flows not covered by automated E2E tests.
Prerequisites:
- iPhone with iOS 17+
- Apple ID for Sign In with Apple testing
- Fresh install of the app (or reset via Settings > Reset Onboarding)
manish
/ manta-v1-limitations.md
Last active
January 12, 2026 19:20
Manta V1 Limitations and Missing Features (from VLDB Paper)
Updated: 2026-01-12
The system operates as a "Single-Node Only" setup without replication, sharding, or distributed queries. Additionally, there is "No Query Admission Control" - meaning no concurrent query limits, memory budgets per query, or queueing mechanisms.
manish
/ cpu_hotspots.md
Created
January 4, 2026 05:23
Manta Query CPU Hotspots - Profiling Results (2026-01-03)
manish
/ benchmark_comparison.json
Last active
December 30, 2025 16:29
Manta Query Optimization Benchmark Report
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "benchmark_info": { | |
| "tool": "ultimate_perf_test --quick", | |
| "iterations_per_query": 20, | |
| "metric": "TrimMean (ms)", | |
| "dataset": { | |
| "total_rows": 100000, | |
| "total_columns": 106, | |
| "core_columns": 12, | |
| "sparse_metrics": 90 |
manish
/ benchmark_comparison.json
Created
December 30, 2025 07:30
Manta Performance Benchmark: optimizations (e1e6505) vs main (c809a5f) - ~8.5% improvement
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "benchmark_info": { | |
| "tool": "ultimate_perf_test --quick", | |
| "iterations_per_query": 20, | |
| "metric": "TrimMean (ms)", | |
| "dataset": { | |
| "total_rows": 100000, | |
| "total_columns": 106, | |
| "core_columns": 12, | |
| "sparse_metrics": 90 |
manish
/ benchmark_comparison.json
Created
December 29, 2025 22:08
Manta Performance Benchmarks: main (c809a5f) vs optimizations (1ecc713) - 36.2% improvement
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "benchmark_date": "2025-12-29T22:07:18.976341494+00:00", | |
| "branches": { | |
| "main": "c809a5f", | |
| "optimizations": "1ecc713" | |
| }, | |
| "dataset": { | |
| "rows": 100000, | |
| "columns": 106, | |
| "min_cols_per_row": 16, |
NewerOlder