EmpCloud — Delivery intelligence

Ranks measured GitHub + Hubstaff signal only — not a performance review, not developer worth.

Generated by an AI model reading commit/PR content, complexity and tracked time. It cannot see DevOps, incident response, mentoring, architecture, code review depth, or work outside the connected repos — read each person's caveats before drawing any conclusion about them.

Conclusion

claude-opus-5 · period Sat Aug 29 2026 00:00:00 GMT+0000 (Coordinated Universal Time) – Mon Sep 28 2026 00:00:00 GMT+0000 (Coordinated Universal Time) · generated 2026-09-28 08:34 UTC · triggered by systemd:ranking · 24/27 people had evidence to rank on · ~$0.819

**The period was heavy on shipping and light on review.** Across the org almost nobody reviewed anyone else's pull requests, and no work unit reached formal acceptance, so every judgement here rests on commit and PR titles, churn, and real AWS activity rather than on a tracker.

**Harsh Patel stood out** for doing both sides of the job well: 30 merged PRs with clear, phased titles plus AWS changes on 18 separate days. Rushendra Varma Mudunuru wrote no code at all but ran identity, access and quota work on 21 distinct days, which is the most sustained infrastructure evidence in the set. Pankaj Kumar and Anshuman Parmar also paired real code with real deploy and DNS work.

**The gap to watch** is people with full tracked months and no visible output: mohammad saad, Anmol Music and Rushikesh at the low end of code detail. That is usually a mapping problem, not a work problem, but it needs checking. Several very large commits in this set are repo backups or branch merges, not authored code, and are called out individually.

Ranked (GitHub + Hubstaff evidence, evaluated by Claude)

Rank reflects measured delivery signal in this dataset, not overall developer worth — read the caveats.

#	Name	Tier	Confidence	Rationale	Caveats
1	Harsh Patel	top	0.55	The strongest evidence on both surfaces. 31 PRs opened and 30 merged at an average difficulty of 4.5, with titles that read like planned delivery rather than filler: "Add matchday Phase 2 durable payments, outbox worker, and EB deploy." (PR 53), "Add Phase 4 offline POS foundation and fix location catalog cart merg…" (PR 57) and "Phase B: Coritiba branding, exit gaps, gate integration, and security hardening" (PR 62). On AWS he was active on 18 distinct days across 19 services, which is real sustained ops work and not a single burst. Code side matches, for example the large "i18n (EN + PT-BR) Frontend" change in NexusAI. coxa-1touch #53 coxa-1touch #57 coxa-1touch #62 coxa_crm #1 NexusAI@0bd983e	1277 infra events is by far the highest count, but the sample shows the same DeleteSecurityGroup on the same security group repeated many times, so the count overstates the amount of distinct infra work. The 18 active days, not the event count, is what earns this rank. No reviews given or received on any of the 31 PRs, so nothing here was independently checked by a teammate. No connected task tracker, so no one can confirm these PRs were the work that was actually asked for. Infra work is concentrated in the 1touch Cloud Infrastructure account (1237 of 1277 events).
2	Rushendra Varma Mudunuru	top	0.40	Zero commits and zero PRs, but the most sustained infrastructure record in the team: 289 write events on 21 distinct days across 17 services, against a team median of 1.5 active days. The event names are the careful kind, not repetition: DeleteUser, RemoveUserFromGroup, DeactivateMFADevice and DeleteLoginProfile to offboard an account cleanly, CreateAccessKey and TagUser to onboard another, RequestServiceQuotaIncrease and a support case, plus ModifySecurityGroupRules. Under the current policy this is top-tier contribution on its own.	153.4 tracked hours all landed on Thunder Marketing's Project, which is not mapped to a repo, so we cannot connect the time to the AWS work directly. Account and access administration often comes with approvals, tickets and conversations that leave no CloudTrail write, so the visible events are likely the smaller part of the job. Almost all work sits in the 1touch Cloud Infrastructure account. No code evidence at all this period, so nothing can be said about code quality either way.
3	Pankaj Kumar	top	0.55	Solid delivery on both sides. 83 commits with real scope, including "Add isolated TheSportsDB proxy under /api/sportsdb." in Brazil_CMS and a matching Mongo-cached proxy in cms, plus site integrations like "Implement RemoteVozApi for editorial content and enhance sports data …". On AWS he was active on 20 distinct days, and the sample is the launch path for those same sites: CreateDomainAssociation and UpdateApp on Amplify with matching ChangeResourceRecordSets in Route53. Code and infra line up, which is the best sign the evidence is honest. Brazil_CMS@a20ec6c cms@26be58b lavozec.com@5d62e02 puntoecuador.world@1bce1ac	1531 infra events is the highest raw count in the team and much of it is repeated DNS record writes, so read the 20 active days rather than the event total. Opened no PRs at all, so none of this code went through even a nominal review gate. Some commits are scaffolds, for example "Initial commit from Create Next App", which inflate churn without much thought behind them. Work spans three AWS accounts, mostly 1touch Cloud Infrastructure.
4	Anshuman Parmar	top	0.55	13 PRs opened and 11 merged on POSITION_INTELLIGENCE, and the small ones are the most convincing: "fix(collector): claim due sources atomically (stop duplicate fetches)" (PR 18) and "fix(intel-worker): revert duplicate cluster_pending beat schedule (hotfix)" (PR 23) show someone who understood a concurrency bug and a bad schedule and fixed them precisely. Bigger work is there too, with PR 19 adding 10-state sources, source-health APIs and failure tracking. Infra is real deploy work: Elastic Beanstalk UpdateEnvironment and CreateApplicationVersion with autoscaling SuspendProcesses and ResumeProcesses around it, on 12 distinct days. POSITION_INTELLIGENCE #18 POSITION_INTELLIGENCE #23 POSITION_INTELLIGENCE #19 POSITION_INTELLIGENCE #11 POSITION_INTELLIGENCE@123d614	The 91,526 total churn is dominated by integration merges, for example PR 11 at 59,927 churn and a 162,235 churn snapshot commit. That is moved code, not written code, so do not read the churn as effort. PR 19 and PR 20 were never merged, so that collector work may still be unfinished. No reviews on any PR, including two that carry 8 difficulty points. 280 of 306 infra events sit in "account 1" and many are GetSessionToken calls, which is routine credential use rather than a change.
5	Aziz Bohra	top	0.55	Three PRs, all merged, all scored at the top of the difficulty scale, and all clearly described: "feat(squadb): trust gates, provenance, volume rollup, and CI hardening" (PR 16) and "feat(squadb): Day 3 news path (near-dup to story to incident)" (PR 21). The email-service commits carry the same weight, for example "prove mailbox recovery reconciliation" and "integrate auth readiness controls". Infra backs it up on 12 distinct days with CreateEmailIdentity, PutEmailIdentityDkimAttributes and the matching Route53 records, which is exactly the domain setup that email service needed. POSITION_INTELLIGENCE #16 POSITION_INTELLIGENCE #21 POSITION_INTELLIGENCE #17 email-service@9b23c87 email-service@63ffd78	203 commits is over five times the team median of 39 while only 3 PRs were opened, which suggests most work went straight to a branch without a review gate. Check whether the commit stream is being split finer than the logical changes. 26 docs and 12 chore commits are a sizeable share of the total. All infra work is in the 1touch Cloud Infrastructure account and is heavily DNS and SES, so it is narrow in kind even if spread over many days. No reviewer looked at the three 8-point PRs.
6	Ankur Kumar	solid	0.55	134 commits with substance behind most of them: "migrate Nexus storage to AWS S3", "neighborhood hub matching, empresas pagination, claim flow order," and "add admin endpoint to hide test data" all describe a specific problem being solved. Infra is steady rather than showy, with 102 write events on 14 distinct days, mostly ModifySecurityGroupRules, which is ordinary but real network hygiene. NexusAI@db85c53 coxa-hub@db88753 coxa-hub@a6f6246 coxa_crm@8feb822 NexusAI@4b0162b	Two of the top commits are scaffolds ("Initial commit from Create Next App", "Initial commit from Create Twenty App") that together account for over 10,000 churn with no authored work in them. Two NexusAI commits have identical titles and identical 2,616 churn, which looks like the same change committed twice. No PRs opened, so no review on any of it. 248 tracked hours is the highest in the team and is spread across seven projects, but hours alone say nothing about output and are not counted in this rank.
7	Abhishek Kulkarni	solid	0.55	131 commits with genuinely large, well-named feature work: "implement complete autonomous agentic system (§14) with…", "Complete Persona Fusion, Classification Pipeline, and Political…" and "add provider-neutral deployment manager". The fix "close finished mixed Ask jobs and stop stale Need…" reads like a real behaviour bug being closed out. Infra is small but purposeful: CreateBucket for a database backup bucket, CreateKeyPair and RunInstances to stand up a host, on 5 distinct days. POSITION_INTELLIGENCE@2ff2056 POSITION_INTELLIGENCE@f8549ec ZeroHumanMultiAgentV3@2a540a3 ZeroHumanMultiAgentV3@e8f2c50 PulseDesk-Test@35f4be9	No PRs opened or reviewed, so 131 commits went in with no second pair of eyes. Several commit subjects are truncated in the data, so scope has to be inferred from the visible part. 43 of the 51 infra events are in "account 1" and are console logins and MFA checks, which are sign-ins rather than changes. The real infra work is the eight events in the 1touch Cloud Infrastructure account. All 163 tracked hours went to projects with no repo mapping.
8	Rishav Raj	solid	0.55	Only 15 commits, well below the median of 39, but every one is a feature and the titles are specific and sensible: "harden one-time schedule execution and retrieval", "add extras, lineage tracking, and failure snapshots" and "migrate SMA workflows and add follower bulk enrichment". This is the profile of someone doing a few hard, well-scoped things rather than padding the count. The one PR, "feat(followers): add stale bulk-job recovery", is a sensible reliability fix. POSITION_INTELLIGENCE #27 POSITION_INTELLIGENCE@5753ad9 POSITION_INTELLIGENCE@b48e0e7 POSITION_INTELLIGENCE@a78df0d	PR 27 was not merged, so that recovery work may not have landed. "complete migration phases" carries 14,577 churn on a vague title, which makes the scope hard to check. No AWS identity is mapped, so infra shows zero. That is not evidence he did no infra work. 194 tracked hours with 15 commits is a wide gap. The likely explanation is unmapped work, debugging or unmerged branches, but none of that is visible here.
9	Akash Kumar	solid	0.55	Real product work on email-service and POSITION: "Day 3: audit-log + API-key screens, DNS template generator, DNS retry…" and "complete day 6 delivery controls" are concrete and reviewable in intent. The infra record matches that work closely, with CreateEmailIdentity, DKIM attribute writes and Route53 record changes across several hosted zones on 4 distinct days, which is the sending-domain setup the email service needed. server-codebase-backup@0957d77 server-codebase-backup@df0b34d email-service@ca80d6c email-service@7978ad7 POSITION_INTELLIGENCE@273157d	The two largest commits by far, "Initial private codebase backup from OpenClaw server." at 355,307 churn and "Add remaining projects to server backup." at 595,458 churn, are file dumps, not authored code. Ignore that churn entirely. No PRs opened, so no review on anything. Infra activity is only 4 distinct days and is almost entirely DNS and SES in one account, so it was a short setup burst rather than sustained ops. Only 0.36h of 177 tracked hours landed on a mapped project, so time cannot be tied to the work.
10	Jitu Prosad Saha	solid	0.55	43 commits, close to the median, but the content is good: "complete SMA Instagram and Twitter profile scrapers", "add production backups and monitoring skeleton" and "add inbound pipeline reconciliation and evidence harness" all describe work that someone thought about. "Add scraper execution logs and harden IG/TW followers list-ready waits." is the kind of fix that only comes from actually watching something fail. Small but real infra: SES identity creation and Route53 records, plus setting up his own MFA. SMA_Backend@4582fcb SMA_Backend@a8b4c82 email-service@a3f357b email-service@9dcaa85 email-service@23c8f0c	Two commits titled "add production backups and monitoring skeleton" with near-identical churn look like the same change committed twice across repos. No PRs opened or reviewed. Infra is only 13 events on 2 days, so this is not an ops-heavy record. Only 0.39h of 166 tracked hours is on a mapped project.
11	Dhiraj Ambekar	solid	0.55	43 commits at the median, with a clear front-end and agent focus: "bulk edit UI for personas, bots, and campaigns", "Agent Orchestration console and Honduras boards UI" and "align LangGraph deps and restore bridge auto-pivot approval". Infra is modest but meaningful and spread over 7 distinct days, including an Elastic Beanstalk UpdateEnvironment with a CloudFormation UpdateStack and autoscaling suspend and resume around it, which is a controlled deploy rather than a poke at the console. Social_Media_Automations@8392093 SMA_Backend@085b863 SMA_Backend@072ec7c Social_Media_Automations@4e5c0bd ZeroHumanMultiAgentV3 #1	One commit reverts another in the same sample ("Revert \"feat(agent): wire live narrative pivot control and POSITION i…\""), so some of that churn was undone. The single PR opened was never merged and scored only 2 difficulty points. Some titles are vague, for example "update the current code of dummy number" at 1,482 churn. All 154 tracked hours went to projects with no repo mapping.
12	Rushikesh Powar	solid	0.40	65 commits, mostly CMS API integration work across many sites, with a couple of genuinely engineering-shaped items: "build incident intelligence collection foundation, add MongoDB indexe…" and "Remove legacy PostgreSQL collection runtime". Infra is small but sensible and spread over 8 days, including CreateBucket plus PutBucketEncryption on a new incidents bucket, which shows the encryption step was not skipped. POSITION_INTELLIGENCE@0a3dfbd POSITION_INTELLIGENCE@6a4ca9c EcuadorCentral.com@a3880bb Global_CMS@0409205	Several titles are placeholders or near-placeholders, for example "update" on an 8,055 churn change, which makes the work impossible to check from the log. No PRs opened, so nothing was reviewed. A lot of the commit volume is the same integration pattern repeated per site, so the count overstates distinct problems solved. Repeated StartInstances on the same EC2 instance on four separate days suggests a box that keeps stopping, which may be unlogged firefighting rather than planned work.
13	MOHD NAVED	solid	0.40	Only 19 commits, half the median, but the quality per commit is high and they cluster on hard client-side problems: "queue offline writes, add Club/FanBox/POS PWAs, and fall back S…", "survive API timeouts and brief offline without fake writes" and "cut unnecessary re-renders across club, fan, fanbox, and POS". The offline and timeout work in particular reads like someone who understood the failure mode rather than patching a symptom. Also shipped "DSR cross-system deletion, redaction, and cloud basel…", which is privacy plumbing. coxa-1touch@cf96cab coxa-1touch@ff64930 coxa-1touch@76b11c5 coxa-1touch@c8b12b4	No PRs opened, so none of this was reviewed. The 3 infra events are only his own MFA and password setup, not infrastructure work. 192 tracked hours against 19 commits is a large gap. Performance and offline work involves a lot of testing that leaves no commit, but that explanation is not visible in this data. Only 1.09h of tracked time landed on a mapped project.
14	Manish Chandrakar	solid	0.40	151 commits, nearly four times the median, spread across a dozen localised CMS sites. The best of them are real: "Add Italy sports hub, harden Serie A APIs, and soft-fail flaky upstre…" shows thought about upstream failure, and "Add production SEO, GEO, and pre-commit lint/build checks." adds a quality gate. Much of the rest is rebranding and palette work repeated per site, which is useful but not hard. global-cms-Italiano@dbe84da ecuadorenvivo.com@1b70e1a global-cms-Korean@4224366 EcuadorCentral.com@a1f9bbb	High commit count with 70 commits in the "other" bucket and titles like "husky test" and "first commit" on an 8,018 churn change. The count is inflated by repetition of the same template across sites rather than by distinct problems solved. No PRs opened, so nothing reviewed. No AWS identity is mapped, so infra shows zero. That is missing data, not proof of no infra work. All 167.8 tracked hours sit on one unmapped project.
15	Pramod Soni	solid	0.40	56 commits, above the median of 39, all on the same theme of article rendering, markdown, FAQ support and SEO metadata across the news sites. Titles are consistently descriptive, for example "Add article routing, markdown rendering, and FAQ support" and "Add support for SEO metadata and FAQ handling in articles". It is honest, steady work, but it is the same change applied repeatedly rather than a hard problem. ecuadorenvivo.life@d71b63d elreporteec.com@4c27bd9 global-cms-veitanamese-new@a993ee5 ecuadorfocus.com@1ff5e2a	The commit count comes largely from repeating one pattern across many sites, so it should not be read as more output than peers nearer the median. No PRs opened, so nothing reviewed. No AWS identity is mapped, so infra shows zero rather than confirmed absence. 122 tracked hours is the lowest of the full-time group, though hours are not counted for or against rank here.
16	1Touch Dev	solid	0.07	15 commits, all on Marketing-Sponsorship-Automation, and the content is unusually user-focused: "real server-side pagination for Companies, Proposals, Campaigns …", "quick-win UX audit items, internal metadata leaks, duplicate UI…" and "sidebar imbalance and double-active-highlight bug (UX audit)". The metadata leak item is a genuine correctness and privacy catch, not cosmetics. Low volume, but the work reads as considered and follows a stated audit. Marketing-Sponsorship-Automation@661ea82 Marketing-Sponsorship-Automation@857adc9 Marketing-Sponsorship-Automation@41a99c5 Marketing-Sponsorship-Automation@8e94464	Zero tracked hours and zero infra events, so this looks like a shared or automation account rather than one person. Attribution should be confirmed before this rank is used for anything about an individual. Confidence is 7 percent, the lowest band, because there is almost nothing besides the commits themselves. No PRs opened, so nothing was reviewed. Three commits are categorised as docs or audit write-ups rather than code changes.
17	Pranav Ojha	solid	0.40	35 commits, just under the median, but four of the five largest are titled only "initial commit" on new site repos, which is scaffolding rather than authored work. The genuinely useful entries are "Dev Console created for intelligence ops" on POSITION_INTELLIGENCE and "correction based on audit report", which shows audit follow-through. POSITION_INTELLIGENCE@0ee7869 globla-cms-Indonesian@a2d752b lavozec.com@2c1c466 informeec.com@69d8ff7	32 of 35 commits fall in the "other" bucket and the titles are mostly generic, for example "markdown enhancement + seo implementation" used on more than one repo. This makes the work hard to verify from the log. Very high churn from repo scaffolds should not be read as effort. The 3 infra events are only his own password and MFA changes, not infrastructure work. All 135 tracked hours sit on one unmapped project, and no PRs were opened.
18	Cursor Agent	needs-attention	0.07	Six small commits on email-service, all tightly scoped and clearly named: "Add manual domain DNS verification", "Keep manual verified domains on public DNS checks" and "Treat representative mailbox proof as inbound ready". These read as careful follow-ups in one domain-verification flow. Total churn is under 900 lines, so the contribution is real but small. email-service@b9bc59d email-service@441362f email-service@dacd999	This is a tool or agent account, not a person. The output should be credited to whoever drove it, and identity mapping needs fixing before this line means anything. No tracked hours, no PRs and no AWS identity, so there is nothing else to weigh. Confidence is 7 percent, the lowest band.
19	Claude Code	needs-attention	0.07	One commit in the whole period, "Fix AttributeError crashing entire Gemini structuring batches" at 44 lines changed. It is a good, precise bug fix title and the impact sounds real, but a single small change is all the evidence there is. Football_Scraping_Main@5f168e4	This is a tool account, not a person. Credit belongs to whoever ran it and the identity mapping should be corrected. No tracked hours, no PRs, no AWS identity. Confidence is 7 percent, the lowest band.
20	mohammad saad	needs-attention	0.40	175.5 tracked hours on Thunder Marketing's Project and no commits, no PRs and no AWS events at all. There is nothing in this dataset that shows what the month produced. That is a question to ask, not an accusation, because the most likely causes are an unmapped GitHub account or work that does not land in these systems.	No GitHub identity appears to be mapped, which would make all code work invisible here. No AWS identity is mapped either, so any ops work is also invisible. All hours sit on an unmapped project, so we cannot even tell which initiative the time served. Design, QA, research, support and coordination work would produce exactly this pattern and none of it is visible.
21	Anmol Music	needs-attention	0.40	135.4 tracked hours, all on a project called "Music", with no commits, no PRs and no AWS activity. Nothing in this dataset describes the output. Given the project name this may well not be a software role at all, in which case this ranking system is simply the wrong lens.	Role may be non-engineering, in which case absence of commits is expected and not a problem. No GitHub or AWS identity appears to be mapped. The single project is unmapped to any repo, so time attribution gives no clue about deliverables. Any creative, content or production output would be completely invisible to this dataset.
22	Srija Bandapally	needs-attention	0.40	16.1 tracked hours on Thunder Marketing's Project and nothing else: no commits, no PRs, no AWS events. That is roughly two days of time in a month-long window, so there is very little to read either way.	This person appears twice in the data under two employee ids, so the record may be split and incomplete. No GitHub or AWS identity appears to be mapped. Low hours may reflect part-time status, leave or a start or end date inside the period, none of which this dataset shows. Onboarding, shadowing and meetings leave no trace here.
23	Shakir Khan	needs-attention	0.40	One tracked hour on the CMS project, with a few other projects logged at zero. No commits, no PRs and no AWS activity. There is almost nothing here to evaluate.	One hour in a month strongly suggests a partial start, a departure or an account that is barely in use, rather than a full month of low output. No GitHub or AWS identity appears to be mapped. The 98 percent focused project ratio is computed from a single hour and means nothing at this sample size. Any onboarding or handover conversations are invisible here.
24	Srija Bandapally	needs-attention	0.40	1.5 tracked hours on "Strategy & Financial Consultant" and no code or infrastructure evidence at all. This is the smallest evidence footprint of anyone ranked.	This is almost certainly a duplicate record for the same person as employee 17 and the two should be merged before any conclusion is drawn. The project name suggests a non-engineering engagement, where commits would never be expected. No GitHub or AWS identity appears to be mapped. Advisory and analysis work leaves no trace in any of these three systems.

Not ranked — no evidence this period

Listed, never guessed. Ranking someone here would fabricate a comparison the data can't support.

Cursor Agent, Ruhani Allagh, Ubuntu Paperclip

literally no GitHub, Hubstaff or AWS evidence this period

What this data cannot see

Production incident response that leaves no AWS write: reading logs, restarting something by hand on a box, a phone call at midnight, a Slack thread that saves a launch.
Anyone without a mapped AWS identity shows zero infrastructure events. That is missing data, not proof they did no ops work.
Anyone without a mapped GitHub identity shows zero commits for the same reason, which is the likeliest explanation for the people here with full hours and no code.
Mentoring, code review done verbally, pairing, onboarding and unblocking other people. Review counts are zero across the whole team this period, which almost certainly means review is happening somewhere this system cannot see.
Architecture and design decisions, written specs, and the choice not to build something.
Work sitting in unmerged branches or on local machines, and work in repos not connected to this platform.
No task tracker is connected, so nothing here confirms that what was built was what was asked for, or that it was accepted by anyone.
Hubstaff project names are mostly unmapped to repos, so tracked time cannot be tied to specific deliverables for most people.
Non-engineering output entirely: design files, content, QA test runs, customer calls and analysis.