If you think the job market is tough for job-seekers, have some sympathy for the folks who’ve got to fill the positions that are out there They have to wade through an ever-deeper pool of applicants to find the exact match for ever-more exacting requirements by hiring managers. Human resources personnel who find themselves in this position might want to check out some new semantic web-based services from Monster.com.

“The breadth of jobs to which a job seeker potentially applies is markedly higher today,” says Javid Muhammedali, director of product management at Monster.com. “Six people compete for every opening. The HR staff is vastly outnumbered. When we did our focus groups and studies the recurrent theme was not that they wanted 100 resumes but they wanted the right two resume to be visible right way.” Many HR departments have faced their own job cuts, so there are fewer people to do the live probing of large pools of candidates.Monster thinks its new Power Resume Search, developed through its acquisition in 2008 of Trovix Inc, and its patented, semantic search technology, is just what they need now, Muhammedali says its technology can cut by 65 percent the average four to six week sourcing time in large enterprises. “In a tough economy, when you do need to hire it’s urgent – they’re desperate and they want to hire now.”

The shortcomings of most other employer job-searching tools, he says, come in large part because of the requirements of using complex Boolean logic – it’s hard to construct a truly effective Boolean query and relying on keywords rather than meaning, and the rankings that are based on the placement of those keywords rather than context, doesn’t quickly deliver the most appropriate candidate results. Someone searching for candidates to fill a director of business development position, perhaps with specialty in enterprise software applications, and maybe with a degree from a prominent business school is tricky: Traditional search might miss appropriate candidates who may have held jobs with those responsibilities but under different titles (director of alliances or partnerships, for instance); they may have specified their enterprise software experience in a variety of ways, as expertise with J.D. Edwards or JD Edwards systems, for instance; and it may be a similar challenge to hunt through and specify variations for business schools.

The new Power Resume Search feature tries to meet such challenges by looking at pools of resumes the same way a human would through the use of Its semantic technology. Based on its knowledge bases of job titles, skills, industries and ability it can identify patterns that indicate candidates have appropriate experience, education and so on. It takes the unstructured text resumes submitted by users just as they are and translates them into a structured format that understands the hierarchy of concepts around individuals’ skills, titles, schools, and so on, understanding what words mean on the basis of their spatial relationships to other parts of the resume. The way the engine develops its knowledge of certain resumes involves a significant amount of pre-processing. Before the search is returned in real time it does this to identify the context and tag the contents of the resume to make it easy for retrieval offline. “We can do more complex pattern recognition and natural language processing that our competitors that have to do that on the fly are hamstrung by,” he says..

Monster also has focused its knowledge base to evaluate the length of time a candidate has held a particular job based on the dates presented on the resume, so those with the longest relevant experience can rise to the top. “So, if a candidate says they worked at a certain job at a company for five years we can differentiate that skill from someone with fewer years of experience,” Muhammedali says. “We look at things as a recruiter does: This person was at this job and used these skills, then we look at how many years he worked and what he did in the next most recent jobs. We give the highest weight to the most recent experience if it is of sufficient duration, but as an example if it’s less than two years the next previous job gets the most credit.” It also can pull out of resumes other items it tags as skills to add up candidates’ experience across multiple jobs over the course of a career to give a picture of how much experiences they have in a particular skill.

HR pros can also use the tool to more easily find well-qualified entry-level help, narrowing down their search to present candidates with one or two years of experience to surface before those with 15 years, for instance. “Sales and customer service positions are high on the number of jobs in Monster,” he says. “And it’s common to want to find entry-level sales people.” One customer was looking for an entry level salesperson in its NY office with online media experience, he relates – who also has shown success through meeting goals or winning awards. “So those are concepts, not keywords among their skills,” Muhammedali says. “We’re looking for patterns on the resumes indicating this person is a top-notch sales representative – for instance, that he was a top ten sales producer each month. All things indicative of gaining lot of sales momentum are matched and we have pattern identification indicating what matches that. It’s not a bag of words matching.”

Another example of its capabilities, Muhammedali says, is shown by a large aerospace defense contractor that was searching for an aerospace engineer with experience in the U.S. military and who went to a top school. The engine immediately surfaced an MIT PhD who had been a lead flight test engineer in the military – that candidate used the word aeronautics and astronautics on his resume, and though he clearly had a wealth of aerospace experience, that word never showed up on his resume. If a keyword search had been done on the word aerospace, he would never have shown up. “Several technologies can do a metasearch of multiple databases at the same time, aggregate the results and then do some sort of semantic ranking from the results that do come back,” he says. “But because they are relying on Boolean for the underlying retrieval there is always the possibility they will miss qualified candidates like this one for the aerospace engineere job.”

The new service was launched this quarter. Existing customers can continue to buy the classic Boolean search license through their representatives, but Monster.com is on track to migrate clients to Power Resume Search. It’s the only option that can be purchased online at this point. The costs are about 30 percent higher than classic licenses. Monster also offers a Candidate Spotlight service to those that haven’t subscribed to Power Resume Search; this lets someone posting a job automatically match candidates to it in the background and then presents that user with an instant slate of candidates. “Because we have the candidate DNA view we can show sufficient information so the recruiter can determine if they want to look at the resume.” In other words, they get some upfront information to help them make a purchase decision that targets only relevant resumes.

Is the Power Resume Search service good news for job seekers who want to heighten their chances of exposure to employees? One benefit Muhammedali says he sees is that the semantic technology now enables jobs to stay live for the right audience on the site longer, rather than being quickly replaced as they are in services such as Craigslist where volume rules and good jobs disappear off the home page quickly. “This technology lets your job stay alive to the right seeker,” he says.

Muhammedali also says that Monster.com is planning to bring similar benefits to them: Right now it has in beta a power search for job hunters that uses the same engine Power Resume Search does for employers but with a simplified interface. He says the new service will make it easier for prospects to hone in on the jobs they want – for instance, it understands what “business development” means so they’ll no longer be surfaced every job that has business and development in its qualifications but may in fact be very far from the actual business development management positions the candidate really wants. “In Power Search because it’s more targeted the same search gets you more targeted results,” he says. Instead of getting 500 jobs you may get just seven, but they’re all excellent matches, he says. “So as a seeker you don’t waste time, If you have an email agent set up you will get emailed jobs that do match so you can act on them immediately.”