Jvalin​ ​Dave 
682.215.5246​ ​|​ ​jvalin17@gmail.com​​ ​|​ ​Philadelphia,​ ​PA 
LinkedIn:​ ​linkedin.com/in/jvalindave  GitHub:​ ​github.com/jvalin17  AngelList:​ ​angel.co/jvalin-dave 
TECHNICAL​ ​SKILLS 
 

● Proficient:​ ​Python,​ ​Flask,​ ​Pandas,​ ​Scikit-learn,​ ​NumPy,​ ​MLib,​ ​C,​ ​SQL,​ ​MySQL,​ ​AWS​ ​(EC2,​ ​S3,​ ​Lambda,​ ​SWF),​ ​Git 
● Exposure:​​ ​PySpark,​ ​Hadoop,​ ​Java,​ ​Ruby​ ​on​ ​Rails,​ ​D3.js,​ ​C++,​ ​Jenkins,​ ​Heroku,​ ​PostgreSQL,​ ​JavaScript 
 

WORK​ ​EXPERIENCE 
 
HeavyWater​ ​Inc.​​ ​|​ ​Machine​ ​Learning​ ​Engineer​ ​|​ ​Philadelphia,​ ​PA Apr.​ ​2017​ ​-​ ​Nov.​ ​2017 
● Optimized​ ​datasets​ ​for​ ​machine​ ​learning​ ​algorithms​ ​by​ ​creating​ ​stratified​ ​10-fold​ ​datasets​ ​in​ ​Python, 
achieving​ ​accuracy​ ​of​ ​69%​ ​in​ ​classification​ ​of​ ​documents. 
● Reduced​ ​execution​ ​time​ ​of​ ​optical​ ​character​ ​recognition​ ​(OCR)​ ​algorithm​ ​by​ ​over​ ​50%,​ ​when​ ​replacing​ ​Java 
CLI​ ​with​ ​Java​ ​API,​ ​improving​ ​throughput​ ​by​ ​2%​ ​and​ ​while​ ​reducing​ ​infrastructure​ ​cost​ ​by​ ​7.5%.  
● Implemented​ ​date​ ​extraction​ ​algorithm​ ​using​ ​Java,​ ​resulting​ ​in​ ​improving​ ​accuracy​ ​of​ ​date​ ​extraction​ ​by​ ​17%. 
● Automated​ ​AWS​ ​service​ ​infrastructure​ ​using​ ​Java,​ ​Python,​ ​EC2,​ ​Lambda,​ ​and​ ​S3,​ ​resulting​ ​in​ ​25%​ ​reduction​ ​in 
processing​ ​time​ ​of​ ​insurance​ ​documents. 
● Increased​ ​production​ ​throughput​ ​by​ ​100%​ ​in​ ​2​ ​months​ ​by​ ​debugging​ ​7​ ​workflows​ ​using​ ​Linux​ ​commands​ ​and 
AWS​ ​Lambda​ ​to​ ​transfer​ ​data​ ​from​ ​EC2​ ​instances​ ​to​ ​S3​ ​buckets​ ​in​ ​order​ ​to​ ​avoid​ ​timeout​ ​failures. 
● Led​ ​team​ ​of​ ​3​ ​engineers,​ ​minimizing​ ​overall​ ​cost​ ​of​ ​production​ ​by​ ​20%​ ​and​ ​managed​ ​DevOps​ ​team​ ​based​ ​on 
agile​ ​development​ ​using​ ​CI/CD​ ​tools​ ​to​ ​maintain​ ​stable​ ​release​ ​of​ ​product​ ​on​ ​schedule. 
 
Lumidatum​ ​Inc.​​ ​|​ ​Machine​ ​Learning​ ​Engineer​ ​|​ ​Dallas,​ ​TX Nov.​ ​2016​ ​-​ ​Feb.​ ​2017 
● Designed​ ​self-learning​ ​recommendation​ ​engine​ ​using​ ​Python,​ ​Pyspark,​ ​MariaDB​ ​and​ ​AWS​ ​S3​ ​resulting​ ​to​ ​5% 
increase​ ​in​ ​sales​ ​through​ ​providing​ ​better​ ​targeted​ ​products​ ​to​ ​consumers. 
● Engineered​ ​data​ ​pipeline​ ​through​ ​daily​ ​extraction​ ​of​ ​30,000​ ​lines​ ​of​ ​text​ ​from​ ​log​ ​files​ ​utilizing​ ​Python​ ​and 
MariaDB,​ ​automating​ ​system​ ​to​ ​optimize​ ​preprocessing​ ​time​ ​by​ ​over​ ​67%. 
● Analyzed​ ​and​ ​designed​ ​new​ ​machine​ ​learning​ ​algorithm​ ​to​ ​calculate​ ​lifetime​ ​value​ ​(LTV)​ ​of​ ​a​ ​customer​ ​using 
Python,​ ​D3.js​ ​and​ ​Matplotlib,​ ​resulting​ ​in​ ​40%​ ​more​ ​accurate​ ​prediction. 
 

EDUCATION 
 
Master​ ​of​ ​Science​ ​in​ ​Computer​ ​Science,​ ​University​ ​of​ ​Texas​ ​at​ ​Arlington May​ ​2016 
Bachelor​ ​of​ ​Engineering​ ​in​ ​Computer​ ​Engineering,​ ​University​ ​of​ ​Pune Aug.​ ​2013 
 

SIDE​ ​PROJECTS 
 
Healthometica​ ​|​ ​Software​ ​Architect​ ​|​ ​live May​ ​2016 
Self-monitoring​ ​health​ ​system​ ​for​ ​University​ ​of​ ​Texas​ ​students​ ​to​ ​track​ ​effective​ ​remedies​ ​for​ ​treating​ ​daily​ ​issues. 
● Architected​ ​platform​ ​using​ ​UML​ ​diagrams​ ​and​ ​analyzing​ ​technologies​ ​such​ ​as​ ​Java,​ ​JavaScript,​ ​MySQL,​ ​JDBC, 
and​ ​Lucidchart​ ​for​ ​prototyping​ ​to​ ​decide​ ​on​ ​most​ ​efficient​ ​tools. 
● Designed​ ​schema​ ​for​ ​MySQL​ ​database​ ​and​ ​JavaScript​ ​UI​ ​with​ ​sketching​ ​and​ ​surveying​ ​40+​ ​university​ ​students.  
 
E-mail​ ​Spam​ ​Detection​ ​|​ ​Machine​ ​Learning​ ​Engineer​ ​|​ c​ ode Dec.​ ​2015 
Machine​ ​learning​ ​algorithm​ ​trained​ ​to​ ​analyze​ ​inbox​ ​emails​ ​to​ ​determine​ ​whether​ ​they​ ​are​ ​important​ ​or​ ​spam​ ​messages. 
● Engineered​ ​e-mail​ ​spam​ ​detection​ ​algorithm​ ​using​ ​Python​ ​and​ ​MySQL,​ ​resulting​ ​in​ ​82.6%​ ​average​ ​accuracy. 
● Built​ ​training​ ​and​ ​testing​ ​data​ ​using​ ​random​ ​function​ ​to​ ​select​ ​rows​ ​to​ ​ensure​ ​split​ ​is​ ​unbiased​ ​using​ ​Python. 
● Determined​ ​most​ ​efficient​ ​algorithm​ ​and​ ​built​ ​Naive​ ​Bayes​ ​Classifier​ ​on​ ​training​ ​data​ ​of​ ​4,000+​ ​emails​ ​from 
the​ ​ground​ ​up​ ​without​ ​the​ ​use​ ​of​ ​machine​ ​learning​ ​algorithms​ ​in​ ​Python. 
 
Yelp​ ​Dataset​ ​Challenge​ ​|​ ​Machine​ ​Learning​ ​Engineer​ ​|​ l​ ive​​ ​|​ ​code May​ ​2015 
Predictive​ ​natural​ ​language​ ​processing​ ​(NLP)​ ​in​ ​text​ ​mining​ ​for​ ​locating​ ​most​ ​commonly​ ​used​ ​terms​ ​in​ ​Yelp​ ​reviews. 
● Gathered​ ​over​ ​1.8​ ​million​ ​reviews​ ​utilizing​ ​Yelp​ ​API​ ​and​ ​chunking​ ​mechanism,​ ​refining​ ​reviews​ ​based​ ​on 
industry​ ​and​ ​place,​ ​and​ ​generating​ ​training​ ​and​ ​testing​ ​data​ ​for​ ​machine​ ​learning​ ​algorithm. 
● Applied​ ​Naive​ ​Bayes​ ​Classifier​ ​in​ ​Python​ ​to​ ​build​ ​model​ ​to​ ​predict​ ​average​ ​rating​ ​with​ ​54%​ ​accuracy. 
● Discovered​ ​most​ ​probable​ ​terms​ ​occurring​ ​in​ ​5-star​ ​Yelp​ ​reviews​ ​and​ ​their​ ​distribution​ ​over​ ​50​ ​million​ ​words.