TinyGambe

Julius Gambe

@TinyGambe
GitHub Profile
collaborative and conversational
A collaborative and pragmatic reviewer who focuses on practical solutions and acknowledges context. Often reviews their own code and provides detailed explanations for decisions, showing a willingness to iterate and improve based on feedback.
164
Comments
96
PRs
15
Repos
84
Avg Chars
2
Harshness

Personality

Pragmatic and solution-oriented Self-reflective (reviews own code frequently) Context-aware and considerate of project constraints Open to feedback and willing to iterate Detail-oriented but not overly pedantic Collaborative and team-focused Honest about mistakes and limitations Practical over perfectionist

Greatest Hits

"LGTM! Thank you for cleaning it up also."
"Thanks for cleaning it up!"
"Whoops, my bad forgot that I removed that"
"Got it. I've resolved it."
"Will make this ready for review once we clean it up."
"ignore it cos the bot keeps deleting the makefile"
"too much changes have been made. will be redoing"
"was just for testing purposes"

Focus Areas

Common Phrases

"I think" "will be" "just" "Thanks for" "LGTM" "Got it" "Will do" "I've resolved it" "my bad" "doesn't have" "should" "when you have" "initially it was" "for now" "oops"

Sentiment Breakdown

neutral
102
positive
21
very_positive
1
questioning
8
constructive
3
critical
2

Review Outcomes

APPROVED
43
DISMISSED
2
CHANGES_REQUESTED
2

Most Reviewed Authors

TinyGambe
103
hongjingzhou
19
SuveenE-TF
10
thakkerurvish
8
Zechereh
8
ayc1
4
lozzle
4
frankfeng98
3
paveldudka
3
zackermax-tinyfish
1

Spiciest Comments

tf-common/#144 [view]
whoops vibe coded test cases
hosting-backend/#85 · app/services/job_results_service.py [view]
the logging doesn't work if we don't do that before the log : (

AI Persona Prompt

You are @TinyGambe, a collaborative and pragmatic code reviewer who focuses on practical solutions over perfectionism. Your review style is conversational and context-aware - you understand project constraints and business needs. You frequently review your own code, showing self-reflection and willingness to iterate. Start comments with phrases like 'I think', 'Got it', 'Thanks for', or 'LGTM!' when appropriate. You're honest about mistakes ('Whoops, my bad') and practical about workarounds ('ignore it cos the bot keeps deleting'). Focus on: error handling, integration challenges, performance implications, and whether solutions fit the existing codebase. You appreciate when people clean up code and aren't afraid to acknowledge when things need rework ('too much changes have been made. will be redoing'). You explain your reasoning thoroughly, especially for implementation decisions, and often mention future considerations ('Will create a linear issue for this'). You're supportive of team members' efforts while being direct about technical issues. Keep comments concise but informative, and don't hesitate to suggest practical alternatives or acknowledge when something 'was just for testing purposes'. You balance being thorough with being pragmatic about delivery timelines.

Recent Comments (137 total)

friday/#454 Change type to fill for agentql_type · pyt/tests/integration/unit_test.py [view]
actually, this is already set by the fixture `browser_service` in `conftest.py`. So I'll just remove this line
friday/#454 Change type to fill for agentql_type · pyt/app/tools/agentql_type.py [view]
From my tests, it didn't. It just tries another tool
friday/#455 Use AgentQL ATree [view]
> I don't fully understand the reasoning behind this change, but the idea makes sense The reasoner wasn't giving the correct value 2025-04-02 because the playwright atree was further split into year, month and day so it was just trying to input 04 or 02 or 2025 so the input was giving errors as that's the incorrect format.
friday/#278 Add browser us [view]
BrowserUse isn't going to be evaluated anymore.
friday/#285 Update graph image to include observer [view]
u r so cool
friday/#269 refractor screenshots · pyt/app/models/browsers/base.py [view]
should we set a global constants for file paths?
friday/#269 refractor screenshots · pyt/app/models/browsers/base.py [view]
i'll merge it first then : )
friday/#263 Human review · pyt/app/agents/observer/observer.py [view]
Had to merge main for the browser issue. Might be @SuveenE-TF @jinyangTF who made this.
friday/#264 Update reasoner prompt [view]
When I was working on Human Review, I tried to incorporate the KG to the planning (https://github.com/tinyfish-io/friday/pull/263). But, I guess I can just change the prompt.
friday/#198 Add external browser · pyt/app/services/browsers.py [view]
Actually, doesn't browser service contain code that deals with the behavior of the browser we want to use? (certain custom logic such as screenshotting etc.) It almost follows playground `BrowserService`. Not sure if in the future we'll have more code that'll be added for this.
friday/#198 Add external browser · pyt/app/services/browsers.py [view]
makees sense, ill make the changes.
agentql-apps/#467 Migrate browser wrap · apps/common/common/utils/browser_wrap.py [view]
Hmm, there are custom implementations such as `page_scroll`, `random_click` etc. that the browser-service library doesn't have. And most of the projects follow this class design so I integrated it to the existing class to minimise changes.
agentql-apps/#167 Tripadvisor browser refractor [view]
> I stopped reviewing at some point due to repetitive nature of my review feedback. Thanks for looking into this. Currently, it's not ready for review as we're iterating fast to deliver the POC so there's a lot of code quality issues. Will make this ready for review once we clean it up.
agentql-apps/#365 fix room capacity issue [view]
LGTM
agentql-apps/#354 refactor retry logic of orchestrator service [view]
Thanks for cleaning it up!