Harbie

La Vida Loca • @ManCity

[Harbie @harbie_]

@_hxmmed Happy birthday G🥳

[Harbie @harbie_]

@_hxmmed @getpostman BYO boss.

[Harbie @harbie_]

@DavidEvbo Happy birthday 🥳!!!

[Tobi @kamteezey]

Okay, I didn’t expect this much reaction. Since I posted this, I’ve gotten a couple of FAQs in my DM, so I’ll try to answer them here directly.

Aside from Goldman, are there other companies that take summer interns?

Yeah, almost every top company you can think of takes summer interns. Other ones I know that have taken interns from Africa include Bloomberg, Meta, Google, Palantir, and JP Morgan. I’ve attached links to their applications website at the end of this tweet.

Are the internships open to Software Engineers alone?

No. Although SWEs tend to be the most common, these companies recruit interns across different disciplines: Production Engineering/SRE, Cybersecurity, Systems Engineering, Product Design, Finance, etc. Check their careers websites for details.

When do applications open?

Usually from mid-to-late August till October, with interviews running into the following year. Goldman’s application opens on August 15, so start prepping. If you start now, you could be in London by this time next summer :)

Can I apply if I don't attend the schools you referenced?

Absolutely. Your school can influence how early you hear about these opportunities, but it doesn't determine whether you can get in. Focus on preparing well and applying.

How much do they typically pay, and do they offer sponsorship?

Back in 2018, Goldman paid SWEs around £47,000 per annum (summer internships are usually 10–12 weeks long, so this is prorated), excluding relocation bonuses. Other top financial firms in London(aside from hedge funds, which are in their own league lol) typically paid around the same, while Big Tech companies generally paid more. The last time I checked, Meta's SWE internship compensation could be as high as £60,000 per annum (also prorated). Check Glassdoor for the most up-to-date figures.

Yes, all of them offer visa sponsorship for both interns and full-time hires, and they fully cover flights and visa-related expenses.

What’s the best way to prepare?

- Get your CV ready. Regardless of what level you're in at uni, you should have a CV, even if the first draft is only a few lines long. Keep improving it strategically as you progress through school.

- Practice Data Structures and Algorithms a lot. Use platforms like HackerRank, LeetCode, and https://t.co/jXd2xRruaw

- Build small side projects and be ready to talk about them during interviews.

- Spend time understanding how large scale software systems are built. Almost every tech company has an engineering blog where they publish technical articles about their systems, so leverage this.

- Get good at using AI and understand best practices. Some companies now allow candidates to use AI during interviews.

- Do lots of mock interviews. Also spend time on each company's Glassdoor page and LeetCode Discuss to review past interview questions.

- Finally, don't self-reject. Apply even if you don't think you're ready. Let the company decide you're not qualified and not you deciding.

Application links:

Goldman Sachs: https://t.co/6asOXsIq3L

Bloomberg: https://t.co/eEkzefJePt

Meta: https://t.co/Wc5OGDd122

Google: https://t.co/hUlw0zQdcW

Palantir: https://t.co/cLMxV76TTW

JP Morgan: https://t.co/aI9husY1ug

[Tobi @kamteezey]

I see a lot of curiosity about this, so I’m going to provide some insight as someone that has similar experience.

Short answer is no. Neither company has a special exchange program with UNILAG. Goldman Sachs in particular had (maybe still has) a dedicated African Recruiting Initiative (ARI). I know this because I was part of the first set of interns recruited under ARI in 2018 while I was an undergrad at LAUTECH. I later returned as a full-time engineer and interviewed students for the same program.

Goldman started ARI in 2018 as an effort to improve diversity at the company. Software engineers flew in from London, New York, and Hong Kong to Lagos to hold a recruitment day where they shared insights about the company and interviewed candidates through group projects and 1:1 interviews, followed by HackerRank assessments and final remote 1:1 interviews. I found out about the opportunity through Stutern and Jobberman at the time.

The inaugural program recruited both full-time engineers and summer interns. Successful interns were hired into the London office, and Goldman partnered with SEO (hence the SEO connection) to prepare candidates for corporate life before they moved to London. Candidates who did well during the internship got return offers as full-time engineers, so many of us already had full-time offers before graduating.

The first cohort was quite diverse, with students from UNILAG, Covenant, LAUTECH, UI, OAU, and Ashesi (Ghana). It was mostly dominated by students from UNILAG and Covenant, largely because of access to information. They found out about the program, shared it among themselves, and prepared together, and it’s the same reason they continue to dominate YoY.

The program has continued since 2018, with subsequent years mostly hiring summer interns, so it’s mainly open to students still in school. There have also been a lot of changes to the process over the years. Applications for the next summer usually open around November/December, so if you’re a student looking to work at these firms, keep an eye on LinkedIn or regularly check their careers page.

Personally, the program was life-changing, and I tried to pay it forward by mentoring a few friends who also got into the program and were retained as full-time engineers.

Many of us from that first cohort have since grown into VPs at Goldman or moved on to become senior engineers at Meta, Google, Microsoft, Palantir, DRW, and Citadel etc while others have gone on to start companies.

So if you’re a student looking to work at these companies, the easiest way in is through the internship pipeline, so hustle for it while you’re still in school. That said, I’ve also seen people who didn’t get in through the internship route come in later as full-time engineers, so that’s very possible too.

Hope this helps someone.

[Harbie @harbie_]

@Jaybaba_ @oreke_lewa @_hxmmed come and defend your city.

[Harbie @harbie_]

Kepano has cooked again 🔥

[Obsidian @obsdmd]

Anything you can do in Obsidian you can do from the command line.

Obsidian CLI is now available in 1.12 (early access).

[Harbie @harbie_]

@_hxmmed @Salmahpopcorn Congratulations to you, yours, her and hers 🎉. May Allah bless your home.

[Harbie @harbie_]

@_hxmmed The second statement is a big lie

[Harbie @harbie_]

@horlaarsco Damn. Bosss 🥳 🔥

[Harbie @harbie_]

@tobyasky @kamteezey this one should be affordable for you.

[Harbie @harbie_]

@kamteezey Very wild wording.

There’s some technical context to the issue here

[Alex Onyia @winexviv]

I promised I will share a detailed technical report of what happened after our review with JAMB core system.

Carefully read it please…

JAMB 2025 UTME Technical Review Report

On Tuesday, the 14th of May 2025, a high-level technical review session was convened at the Joint Admissions and Matriculation Board (JAMB) headquarters in Abuja. The meeting, presided over by the esteemed Registrar, Professor Ishaq Olanrewaju Oloyede, was initiated in direct response to the mass outcry that followed the release of the 2025 Unified Tertiary Matriculation Examination (UTME) results the previous Friday. The objective of this gathering was to unravel the root causes behind the unexpectedly poor candidate performance and to establish clear mitigative measures to restore confidence in the integrity of the UTME assessment process.

The meeting began promptly at 10:00 a.m. and was attended by a distinguished panel of stakeholders. Present were heads of key directorates within JAMB, and Lead Systems Analysts. Also in attendance were delegates from the CBT Centre Regulatory Committee, representatives from the Educare Technical Team, and lead engineers from the consortium of software vendors responsible for the examination engine infrastructure.

Discussions commenced with a comprehensive analysis of the existing system architecture. The panel reviewed the software stack powering the CBT engine, paying particular attention to how examination content was delivered to candidates. An important distinction was made between server-streamed and locally cached delivery methods.

The discussion further investigated the existence and efficacy of randomization mechanisms for both questions and answer options. This scrutiny was aimed at determining whether shuffling protocols were uniformly enforced, how answer permutations were managed, and if consistency of correct answer mapping was maintained across candidates following randomization.

Attention then shifted to the scoring and marking logic. This line of inquiry included assessing the capability of the system to reconstruct question-and-answer mappings and evaluate response accuracy with complete transparency. It was critical to determine if each candidate’s raw responses were stored, retrievable, and auditable.

The team also rigorously examined JAMB’s quality assurance and testing frameworks. This involved reviewing the load testing procedures previously employed, verifying the consistency of deployment builds across centres, and probing whether any last-minute hotfixes or patches were applied during or after the examination period. System logs, particularly incident and error records, were meticulously analyzed to identify any anomalies that could suggest failures in content delivery, timing accuracy, or candidate response capture.

Another focal point of the review was *the potential for human factor influence* . The panel sought clarity on whether any manual post-processing occurred during score collation or validation. The conversation concluded with an assessment of JAMB’s response readiness to Freedom of Information requests and its willingness to publish anonymized candidate-level result data as a measure to enhance public trust.

*One of the most critical discoveries* made during this session revolved around *three* major systemic changes introduced in the 2025 UTME. The first was a shift from the traditional *count-based analysis* to a more robust *source-based analysis* of results. In previous years, JAMB evaluated the integrity of examination sessions primarily by counting the number of responses submitted per session. If the majority of candidates in a session of 250 submitted a near-complete set of answers, the session was deemed valid. Any significant deviation led to disqualification of that centre’s results. However, in 2025, a more advanced model was adopted—one that focused on the actual source and logic of the answers provided, rather than just their quantity.

The second change involved *full-scale shuffling of both questions and answer options* . This ensured that even two candidates sitting in the same session would not receive identical permutations, thereby enhancing test security. The third change was *a series of systemic improvements aimed at optimizing performance and reducing lag* during exam sessions. This was a major policy change that saw the best and highest Obtained UTME score in 15 years. And this would have amounted to a great achievement by JAMB!

While these improvements were technologically sound in theory, a major operational flaw was uncovered during the implementation phase. The system patch necessary to support both shuffling and source-based validation had been fully deployed on the server cluster supporting the KAD (Kaduna) zone, but it was not applied to the LAG (Lagos) cluster, which services centres in Lagos and the South-East. This omission persisted across all sessions until the 17th session, after which the error was discovered and corrected.

As a result, approximately 92 centres in the South-East and 65 centres in Lagos—totalling 157 centres—operated using outdated server logic that could not appropriately handle the new answer submission/marking structure. This affected an estimated 379,997 candidates, whose results were severely impacted due to system mismatches during answer validation.

To verify the scale and accuracy of this issue, JAMB collaborated with the Educare Technical Team, which had gathered response data directly from over 18,000 candidates. After deduplication and filtering, about 15,000 authentic records were analyzed. Of these, more than 14,000 originated from the regions serviced by the unpatched LAG servers, confirming the technical review's findings. Comparative analyses between JAMB’s internal audits and third-party system evaluations revealed significant overlap, reinforcing the conclusion that the affected centres were indeed operating under impaired conditions.

In response to these findings, Professor Oloyede convened a press briefing at 3:00 p.m. on the same day. He formally acknowledged the oversight and issued a sincere apology to candidates and their families. He announced that all affected candidates would be given the opportunity to retake the examination at no additional cost. Furthermore, recognizing the potential scheduling conflict with ongoing SSCE examinations, JAMB had reached an agreement with the West African Examinations Council (WAEC) to ensure seamless coordination of timetables.

Affected candidates were advised to reprint their examination slips by Friday, 17th May 2025, to confirm their revised test schedules. JAMB also released an official communique titled *MAN PROPOSES, GOD DISPOSES*, within which as a Profound, emotional Section: *“Appeal, Appreciation, and Apology,”* which reiterated its commitment to fairness, transparency, and continuous improvement.

This review, conducted with thoroughness and transparency, signifies JAMB's resolve to uphold the sanctity of its examination processes. Going forward, stronger deployment validation protocols and real-time monitoring mechanisms will be implemented to prevent such oversights.
In summary, JAMB opened its systems to independent reviews to restore public confidence and ensure the reliability of the UTME for all stakeholders. And we hereby report, that this incident was neither a system failure nor administrative manipulation, but an outright human error.

~
Educare Tech Team,
As submitted by
Engr James Nnanyelugo

[Harbie @harbie_]

Successful hate watch.

[Harbie @harbie_]

@_sawzeeyy Congratulations Boss. Send cake na. This deserves special celebration. Bubu is no longer president.

[Harbie @harbie_]

If I die today na PCN kill me.

[Harbie @harbie_]

Severance is good if you stop at the first season.

The rest is pure BS!

[Harbie @harbie_]

@hooj___ Happy birthday G 🎉

[Harbie @harbie_]

@adxgun Just because you live a ghetto life doesn’t mean that definition have to apply to others too.

[Harbie @harbie_]

Must be NP hard but not undoable

[Harbie @harbie_]

Thought about this today.

[David @DavidSHolz]

can someone make a realtime voice-to-voice language translation ai yet? pls? what are we waiting for?