Solarwinds Inc. Developer- Accuracy Team in Bucharest, Romania
As an accuracy data scientist your key role is to maximise the number of unwanted messages that are rejected by the filtering products, while minimising the number of wanted messages that are accepted by the filtering products. You will do this by analysing current weaknesses and upcoming threats, developing signatures and rules that address these, and recommending development work for issues that signatures and rules are insufficient to address. You report to the development coordinator.
Creation of rules and signatures
Develop ClamAV signatures; OrangeAssassin rules; DNSBL, DNSWL, DNSYL, URLBL, and URLBL rules; and auto-lock rules. These will be high quality and performant rules that increase the value of the products, which are specifically targeted at both current weaknesses of the filtering system (as identified in analysis of reported false negatives and false positives) and expected future issues (as identified through analysis of expected threat data).
Work with the development process
Take an active part in the planning process for the accuracy team and attend each daily update meeting (including reporting on the previous day’s work, expected work for the current day, and any likely issues).
Resolving urgent issues
When urgent accuracy issues are identified (e.g. via the support team or through reported false negatives / false positives), assist in providing immediate resolution to the issue (and then develop a plan for avoiding the situation in the future). This might involve deploying additional rules or signatures, manually whitelisting / blacklisting IPs, locking problematic submission accounts, or other techniques.
Improve transparency around classification
Ensure that anyone within the Mail PLG is able to easily know how the delivery and submission filters are currently performing, what the most significant threats are at this time, what threats are expected in the near future, and what improvements are currently being worked on.
QA processes for trained message verification
Review a portion of the trained messages that have been verified as quality control. Where there are errors, ensure that those are corrected in the training corpora.
Identify possible new external sources of data to assist in the classification process (e.g. DNSBL or URLBL feeds, SpamAssassin rulesets, IP reputation feeds, honeypot data). Where legal requirements and policy permit, develop and enhance data sharing partnerships. Engage in relevant communities to both improve our classification accuracy and enhance the reputation of the company.
Linux command line proficient
Basic knowledge of how email works
Basic scripting knowledge in either bash or python
Good pattern recognition skills, analytical thinking;
A passion to confront new malware and spam attacks
Ability to process large batches of data (using IMAP)
Proficent in regular expressions (RegEx)
Nice to have:
Familiarity with email related security systems (DNSBL, DKIM, SPF, DMARC etc)
Experience with working with malware
Experience in either Yara, ClamAV or SpamAssassin (all is a bonus)
SolarWinds is an Equal Employment Opportunity Employer. SolarWinds will consider all qualified applicants for employment without regard to race, color, religion, sex, age, national origin, sexual orientation, gender identity, marital status, disability, veteran status or any other characteristic protected by law.
Job ID 2019-12514
# Positions 1
Category (Candidate Search) Engineering
Position Type Regular
Business Group SolarWinds MSP