Following is a guest post from ff Venture Capital winter intern Max Segan, a Colgate senior majoring in computer science, who is starting at Facebook this summer as a software engineer.  From our perspective, Facebook is both our friend (training hundreds of talented future entrepreneurs) and enemy (competing with our portfolio companies for top talent like Max.)  Max obtained approval from Facebook for the post below, which includes sanitized versions of some of the questions he received.

As I’ve just finished the interviewing/hiring process at Facebook, I thought it would be helpful to share some of the information and insights I gained, particularly for our many portfolio companies who are competing with Facebook for talent.  As the recent popularity of the WSJ article How to Ace a Google Interview showed, the hiring process at the top tech companies is the object of a great deal of interest and fairly little information.  I hope I can provide insight for both applicants and companies looking to hire engineering talent.  There are a number of practices that are worth taking away from how Facebook conducts their hiring process, though some of their practices are specifically to deal with their own fairly unique situation.

While Facebook reportedly received a quarter of a million job applications last year, trying to filter that many technical applicants isn’t a common problem.  To help manage this surplus of applicants they created automated programming challenges to filter out clearly unqualified candidates.  These problems are mostly straightforward "CS 101" type problems.  This is likely not the strategy for small companies.  It requires a large number of potential problems so candidates won’t know the questions ahead of time, and it takes a decent amount of upkeep to manage the servers that compile and run the candidates’ code.

After the initial filtering process, Facebook chooses to give candidates up to three phone interviews. They are around 45 minutes each, and they use  Collabedit to see you code in real time.  They generally include only one or two technical questions and about five to ten minutes for the interviewing engineer and the candidate to ask each other questions about work at Facebook and what projects they’re both involved in.  The questions that I got were all pretty standard questions one might find in a "Data Structures" or an "Analysis of Algorithms" class.

The Facebook team was very good at following up quickly; after the first two phone interviews they contacted me regarding the next round the following day.  Following the third interview, they emailed only two hours later asking me to pick a time to fly in.

The final stage of the interview process was a day long visit to the facilities, complete with more technical interviews.  The process included a comprehensive tour and a meal in the dining facility, a technical talk, and then three more interviews.  They followed the same structure as the phone interviews, and the questions were fairly similar.  As with the phone interviews, every question was only finished after code was written up.

Example Question 1 (requires programming knowledge)
Given a list of valid words, return them grouped together by which are anagrams of each other.
[Answers to all questions at the bottom of the post]

Example Question 2 (requires programming knowledge)

Given an array of integers determine if any three sum to 0.  Try to find a solution that uses O(1) extra memory.

My top take aways from a recruiting perspective:
1) Fast turnaround between interviews. In a very competitive market for quality engineers, good candidates are almost certainly interviewing elsewhere, too; it’s important to lose as few as possible due to poor response time or anything else easily preventable.  Keep candidates engaged.

2) In person interviews are a must. If you’re hiring for a full time position it can be very risky to hire after only using email/phone/Skype.  You can’t be certain enough of their ability or their fit, and the cost of a failed hire is incredibly high -- even more so when contrasted with the benefit of a good one.

3) Always ask for an implementation. A lot of people can be very good problem solvers, but an employee who can’t code will be a bottleneck on projects.  This is particularly true for smaller companies.

4) Provide interesting problems. Entice them with the difficult and interesting questions that they’ll face.

5) Network the candidates contacts. Facebook repeatedly asked for any good engineers I knew.  It’s hard to find good engineering talent right now, and if you’ve found one they probably know others.

6) Use algorithm-type questions. It’s important to make sure the candidate has a good knowledge base, and quality computer science/algorithms problems offer a great way to see how smart somebody is while also testing their data structures and algorithms knowledge.  Even non-technical candidates can tackle these problems if presented in an accessible manner; these types of questions illuminate how bright, quick, and creative a candidate is.

Example Question #3 (no technical knowledge required)
You have two identical eggs, and a building that is 100 stories tall (and then solve it generically for an N story building).  You want to determine the highest floor you can drop the eggs from such that they don’t break, and do so in a way that minimizes the worst case number of times you’ll have to drop an egg.  If an egg breaks it becomes unusable, but if you drop an egg and it doesn’t break you can use it again.  There is no random chance in the drops, and the eggs don’t build up damage along the way (i.e. previous drops are unimportant if it remains unbroken).  What systematic way can you go about doing this?

Example Question #4 (no technical knowledge required)
There are 100 prisoners in a line.  They each have either a red or a black hat on.  They are able to see the hats of every prisoner ahead of them, but are not able to see their own hats or the hats of any prisoners behind them (i.e. the 30th prisoner can see the hats of prisoners 1-29, but not his own or those of prisoners 31-100).
The executioner will start at the back of the line and ask each prisoner in turn to guess their own hat color.  All prisoners still in line can hear the guesses (but no information other than simply “Red” or “Black”).  If they get it right they are allowed to live, and if they get it wrong they are killed immediately.
The prisoners have not yet been lined up or given their hats and are trying to decide on a strategy for guessing.  They want to guarantee the survival of as many prisoners as possible, what is the best strategy you can come up with?

Answers to the Questions
Many questions have multiple answers that are all just about as good as one another.  I’ll first say how efficiently the problem can be solved, then give one such way of how to solve it that well.  If solved it as (or more) efficiently then of course it’s not an issue if you used a very different method.

Answer to Question #1
Let m be the average number of characters per word and n the number of words.
Time: O(m*n)
Space: O(n)
The first realization is that words being anagrams is an “equivalence relation”, meaning a word is an anagram of itself, if word A is an anagram of word B then word B is an anagram of word A, and lastly if word A is an anagram of word B and word B is an anagram of word C then word A is an anagram of word C.  This should be pretty clear, and armed with this we realize that to be anagrams just means we have the same number of each letter (so 5 “a”s, 2 “b”s, 0 “c”s, etc).  Thus for each word you can simply count up how many of each letter it has (in big oh notation this is linear in the length of the word), and then hash the resulting twenty-six-tuple, with the value we store being a list of pointers that all have the same number of letters.  The resulting hash map has all anagram groups together.

Answer to Question #2
Time: O(n2)
Space: O(1)
Start by sorting the array which takes O(nlogn) time.  Then repeat the following process for each element starting from the leftmost (smallest element).  Keep the element as the invariant, and for the other two elements that you’re currently looking at select the rightmost and leftmost in the remaining array between the invariant and the end of the array.  If the three you’re currently looking at sum to a positive number decrement the rightmost pointer by one, if they sum to a negative increment the leftmost pointer by one.  This process continues until the pointers meet or we find a set of three elements that sum to 0.  This takes O(n) time for each invariant, of which there are n, leading to the total run time of O(n2).

Answer to Question #3
Find the smallest number, k, such that 1+2+...+k= N.  Then start by dropping it from k, move up by k-1, then move up again by k-2, etc until it it breaks.  Then use the second egg to do a linear walk up from the last place it didn’t break up.  This can take at most k+1 drops.  For the 100 floor building this means 15 drops in the worst case.  The reason to decrease the steps by one at each time is that if the first egg breaks after three drops we want the worst case for the remaining work to require two less drops than if it breaks on the very first drop, otherwise there is some inefficiency we’re not capitalizing.
This takes about 2n drops for the mathematically interested.  (I leave the proof of this to you.  If anybody is extremely interested comment, and I’ll explain in more detail).

Answer to Question #4
Have the prisoner in the back count how many black hats they sees.  If it’s an even number they guesses black, odd red.  This prisoner may or may not die depending on luck.  The next prisoner can use the fact that counting their hat the number of black hats is either even or odd to see if theirs is black (if it is then they should see an odd number of black hats, otherwise the people in front of them should have an even number).  The process then propagates forward and all but the first prisoner is guaranteed to be saved.

For the computer science minded people odd/even is the same as doing bit-wise XOR of the hats (say black is one, and red zero).  It is an invertible function and after the last guy all prisoners will know the XOR of all the prisoners besides him and can by the time they go they will know the values of everybody in front of them and everybody behind to invert the function properly.

facebook photo credit: sitmonkeysupreme

I’ve recently met with several universities, nonprofits, and government employees who’ve all asked the same question: how can we promote entrepreneurship?  Just like me, they’re distressed about the poor economy and jobs situation, which contrasts dramatically with the growth and tight labor market in the technology ecosystem.  They’re concerned with this issue at both the local and national levels.

The first and primary role of a government is to provide basic public goods competently.  Like many Americans, I’m disappointed that our government sometimes fails to do this.  Above all, I’d like our friends in government and the policy world to figure out how the US government can deliver what it’s supposed to but sometimes doesn’t: sound money; an affordable budget; effective education; functioning infrastructure (e.g. airports, less-intrusive airport security); functional legal system (tort reform); logical immigration reform; and a functional tax system (tax simplification).  To Mayor Bloomberg’s credit, he is using his bully pulpit for exactly this purpose, e.g., decrying US immigration policy as ‘national suicide’.  Sadly, he doesn’t control our immigration policy.

There are a range of organizations that help address issues in the technology ecosystem, including: Women in Technology, Technology Policy Institute, New York Technology Council and various other state technology councils, Center for Policy on Emerging Technologies, Information Technology Industry Council, Silicon Valley Leadership Group, and Tech Policy Central. I support their goals of making entrepreneurship accessible to those who have the ingredients to become a successful self-starter. Like virtually the entire tech industry, I am particularly in favor of Startup Visa, which has the goal of stimulating our domestic startup community through acts to keep our foreign-born entrepreneurs in the United States.

Following is a list of some of the ideas I’ve working on (to various extents).  If you see something that piques your interest and want to get involved, please contact me.  I have ranked this list in ascending order of cost (defined as combined time and money).

-       Educating the VC/angel community on the city/state resources available to them to build companies. I’ll likely organize a Harvard Business School Angels of NY event on this topic in 2012.

-       Studying best practices of VCs in supporting portfolio company operational improvement. I am now leading a Columbia MBA team (ex-Mckinsey and BCG) on a study regarding best practices of VCs in supporting portfolio company operational improvement.  This is effectively a sequel to my study of best practices in deal origination, published in the Journal of Private Equity, Harvard Business Review, Institutional Investor, and Business Insider. You can see preliminary results of this value creation study here.

-       Encouraging non-US companies to set up operations in the US. As the former CEO of an Israeli startup with a (modest) US and UK presence, this is an obvious way to create more jobs here.

-       Creating angel groups with other alumni organizations, using HBS Angels as a model, e.g., alumni of major NY schools (Columbia, NYU) and institutions (McKinsey, Goldman Sachs).  I’ve talked informally with some universities about doing exactly this.

-       Creating a for-profit business with the goal of helping students conduct research projects for businesses, so that they’re more connected with potential employers.  I wrote a blog post on this.

-       Organizing events for local mid- to large-company CIOs / CEOs to meet technology startups, in order to help startups get traction and keep their momentum going.  HBS Angels will likely execute this in 2012.

-       Publishing notes on local tech events. New York has an extremely active Meet-up culture; almost every night there are 3-10 tech-focused events.  However, no one systematically provides notes/videos of these events.  Here’s a win-win way to address this: recruit a team of students who would attend these events and publish their transcripts/videos, so that everyone learns more and faster and can network more easily.  I’m already doing this on an experimental basis on my blog.

-       Publishing joint research. See ffvc.com/topics for a full list of topics we want to research.  Part of our strategy as a VC is to do in-depth research on the questions that interest us, in order to make us smarter and to help improve industry best practices.  We typically do this by partnering with graduate student teams.  In particular, we think it would be helpful to the ecosystem if someone were to publish research on who are the most active and most successful New York-area angels, superangels, and early-stage VCs, in order to identify the leaders with whom entrepreneurs should most want to work.  (it is likely that ‘active’ and ‘successful’ are not synonymous.)

-       Creating a master operating checklist for New York entrepreneurs. Visualize a grid, with the standard entrepreneurial steps down the y-axis (ideation, market research, incorporation, etc.), and the major New York industries across the X-axis (internet, food, retail, fashion, etc.).  It would be valuable to have checklists for all of the intersections in this grid.  We’ve created some but not all of these checklists for the tech industry; see our presentation notes and links on our resources page.  Other useful checklists we’ve identified include Biztree, Wickedstart, Goodwin Founders Workbench, American Express Open Forum Crash Courses, StartupToDo , Entrepreneurcountry, and StartupPlays.

-            Organizing angel pitch nights focused on special-interest communities, including industries (healthcare, retail, fashion) and affinity groups (Women, Asian-American, Latino, LGBT, African-American).  I am now talking with some affinity groups in these domains about co-organizing such pitch nights  under the umbrella of HBS Angels.  Our goal is to source and fund great entrepreneurs, regardless of personal or industry background. We are planning a healthcare-focused evening on February 29, and are working on an evening focused on African-American entrepreneurs and on the mobility sector. Details to come.

Here are some much more ambitious and expensive ideas:

-       Setting up a ‘virtual seastead’, perhaps on UN headquarters territory, for high-quality entrepreneurs who do not have US visas. A team in California is already creating a offshore seastead for non-US citizens who want to build companies in the US.  The UN headquarters in New York is technically located on ‘international territory’.  I’m sure there are already some for-profit businesses there, e.g., restaurants.  I wonder if it is possible to set up an apartment building near UN HQ as a virtual seastead?  I’m unclear whether this legal arbitrage is possible. Is anyone else looking into this?

-       Launching a seastead accelerator offshore from New York.

What other policy initiatives should the local tech community, be working on?

(Image by Getty Images via @daylife)

The New York Fundingpost VC and Angel Showcase will be next Tuesday, Feb. 7, 2012. ff Venture Capital will be there.

37 VC Funds & Angel Groups will be exhibiting their firms during a cocktail party setting from 6:00PM to 9:00PM. The event is sponsored by Credit Suisse, WeiserMazars, Morgan Lewis and SocialRadius. More information, a list of the Funds and Registration are here: RSVP

Pizza/soda provided for all attendees.  Pitching includes Gust (formerly Angelsoft) application and pitch coaching from Mac Lipscomb

Panelists (alphabetic order)

• Steve Brotman (@stevebrotman) Managing Partner and Founder at Silicon Alley Venture Partners
• Charlie Kemper (@ckemper) Partner at Revel Partners and Managing Director at ER Accelerator
• Charlie O'Donnell (@ceonyc) Partner at Brooklyn Bridge Ventures
• David Teten (@dteten) Partner at ff Venture Capital

Startup Pitches

1. Farhan Memon - WikiOrgCharts
2. Nihal Parthasarathi - CourseHorse
3. Jason Nadaf - SureDone
4. Jonathan Rosen - Give.bz
5. Julian Barnes - Body Local
6. Frank Harrison - MedCASH
7. Rob Caucci - SpaceSplitter
8. David Baker - Spins.FM

Agenda

• 6:30pm – Pizza, networking and introductions
• 6:45pm – Investor Feedback Forum – 10 startups pitch to an investor panel; the panel provides actionable advice for each pitch
• 8:15pm – Pitch Showdown - audience voting via Twitter; winners announced and prizes awarded
• 8:30pm – Drinks at nearby location

Format

The Investor Feedback Forum and Pitch Showdown is designed to:

• Help early stage startups refine their investor pitch
• Provide helpful, actionable advice on each startups business model from active, early-stage tech startup investors
• Demonstrate to the audience how investors evaluate startup pitches
• Award prizes to the most viable startups, based on audience and panel voting.

The three winning startups are automatically invited to pitch to the NY Angels.

RSVP.

One of our portfolio companies asked me this week, “Do you all have a list of job boards your portfolio companies have had great success with in recruiting high quality engineers?”

The unemployment rate for quality computer scientists/engineers in the major cities right now is zero; everyone who’s not pursuing his/her own startup can easily get a good job.  Therefore, the best candidates are rarely perusing job boards.  Part of the genius and success of LinkedIn is it’s by far the best tool for soliciting passive jobseekers, i.e., those who are currently employed.

I’m a firm believer that the best way to recruit great talent is to make yourself a magnet for people to come to you, instead of the other way around.  In a Dreamit talk I gave on recruiting star engineers, embedded below, I discuss a host of ways to do this.  In a HBS Angels panel I spoke at last week, Chris Fralic of First Round Capital quoted Chad Dickerson of Etsy on this topic. Speaking at a First Round-organized event, Chad said Etsy requires 100% of their engineers to be active in the technical community: give talks, blog, etc.  As a result, they’ve had no problem attracting great talent to Brooklyn.

That said, while largely cumbersome and cluttered with spam, job boards still accounted for nearly 25% of hiring in 2010.  Nathan Hurst made a useful comparison of tech job boards based on cost/post, headhunter/confidential posts, company size, and word frequency.  Accolo uses its database of thousands of placements to identify the most effective job boards for its clients (Disclosure: I'm an Advisor). Social Recruiting Report ranks the top tech job boards by Google Page Rank.  Smashing magazine and Sociable Blog have lists of IT/Tech job boards, but don't do a detailed comparison .    Engineering Recruiting promises to put you in touch with the appropriate engineer recruiter, but I'm a bit skeptical of how effective or worthwhile a service like that is, given the conflicts of interest.

What do you find are the most effective recruiting job boards?

Thanks to Joan Xie for research help.

The Founder Institute recently posted a video from a talk I gave a little while ago on Startup Fundraising Hacks. Although my focus in this talk is raising capital, the ideas and technologies I discuss will also help you in recruiting star developers, sales, and even dating. The talk is dated in one respect: at the time I gave it, AngelList hadn't yet hit the level of visibility that it does today, so I didn't discuss it at all.

How to Raise Capital for Your Startup with Social Media

ff Venture Capital and HBS Angels of New York are happy to be co-sponsors of Flashpoint's upcoming New York Demo Day, Wednesday, January 18, 2012, 2:30 PM to 6:00 PM, held at Union Square Ventures.  Flashpoint is a startup accelerator program at Georgia Tech. Some of their portfolio companies look very promising. The current crop includes:

I hope to see you there! RSVP

I'm one of the judges for an MIT Enterprise Forum of NYC Pitch event on January 31 focused on Digital Commerce companies:

Panel Speakers:
- Eric Nordlander, Engineering Partner @ Google Ventures
- David Teten, Partner @ ff Venture Capital
- Albert Wenger, Partner @ Union Square Ventures
- Sim Blaustein, Venture Investor

Participating Companies:
- Agnik - data analytics company for distributed, mobile, and embedded environments
- Echoer - finalist at the Le Web competition
- Moviecom.tv - Red Herring Global 2011 Winner
- Megawatt Hour - decision-making platform for energy purchasing
- WishMash - The social wishlist platform

Panel Moderator:
- Lori Hoberman, Partner, Chadbourne & Parke LLP

DATE: Tuesday, January 31, 2012

AGENDA:
5:30pm - 6:00pm: Reception
6:00pm - 7:30pm: Business Plan Presentations
7:30pm - 8:30pm: Networking

PLACE:
Chadbourne & Parke LLP - Directions
30 Rockefeller Plaza, New York, NY, 10112

REGISTRATION:

Free to members of MIT Enterprise Forum
$50 non-members, $10 extra at door
All members and guests are welcome.

REGISTER HERE.

We also have a great panel coming up next Thursday night, Jan. 17, on “How Investors Are Increasing Their Returns Through Collaboration and Technology”.   This is another joint event between HBS Angels of NY and the HBS Club of New York.

I hope you can join us Thursday night at a joint event of the Harvard Business School Angels of New York and the Harvard Business School Club of New York, on "How Top Venture Capitalists Create (and Sometimes Destroy) Portfolio Company Value".  I'll briefly discuss my research study on this topic, and then we'll have an all-star panel speaking.

Christopher Fralic is a Managing Partner at First Round Capital’s New York office, and has focused on a number of the firm’s investments in Advertising Technology, Social Media, Ecommerce, Gaming, Mobile and more.  Mr. Fralic has 25 years of technology industry experience, with significant Internet business development roles since 1996.  Prior to joining First Round in early 2006, Mr. Fralic was VP of Business Development at social bookmarking and tagging company del.icio.us through the Yahoo! acquisition.  Mr. Fralic was also one of the early employees and VP of Business Development at Half.com starting in 1999, and after the eBay acquisition, spent six years with eBay in a variety of entertainment, business development, and media roles.  Mr. Fralic earned his BS in Finance from Villanova University and his MBA from St. Joseph's University in Philadelphia.  You can reach Chris at www.nothingtosay.com and @ChrisFRC.

Dave Tisch is Founder and Managing Director of TechStars in NYC and is an angel investor through the BoxGroup.  BoxGroup is an investor in companies like Boxee, Flavors.me, Goodsie, fab.com, art.sy, Coursekit, Skillshare, GroupMe, DataMinr and many more.  You can reach Mr. Tisch at Tisch@techstars.com and @davetisch.