Michael Jagdharry

kdb+ Developer

Talent

New York City, New YorkMember since August 28, 2023

0 followers · 0 following

Hire this person

Send a job offer directly to this candidate

About

Summary of KDB+ skillset/projects

I’ve developed and maintained a large q/kdb+ codebase. The database is 100s of terabytes large and contains nanosecond-scale data back to 2017. I have Improved legacy code and maintained various applications written by previous developers. Such apps include a trading alert system (for example if a trader's session terminated unexpectedly), a q process monitoring system (for example alerting if processes suddenly crashed), and a script that calculates the price for a custom (financial) derivative product, based off treasury interest rates. The script is based on the Newton-Rhapson algorithm to numerically approximate a (mathematical) derivative, in this case the duration of a bond, and ran off a real-time database. These have all required thorough testing and debugging to ensure accuracy and reliability.

My last client is a brokerage; the primary application I supported sends trade reports to each “secondary”-client trading through the brokerage. I have had to communicate effectively with stakeholders and team members to understand business requirements and translate these into a technical q/kdb+ solution. These trade reports contain the volume traded for treasuries with maturities ranging from less than 1 month to 30 years, and the client's ranking (in terms of total volume traded) against others per each treasury type. Such ranking reports are generated with time horizons of the given day, the month to date, and year to date. These reports are automatically generated at end of day and sent out to each client.

The primary application I developed is a partitioned historical database storing the servers' disk read/write speeds and synchronous/asynchronous interprocess communication speeds. This was a flexible application that wrote directly to disk and could be restarted at any point without data loss. The brokerage needed new Linux servers to support their growing operations and had conflicting results for speed tests on these new servers. Thus I was tasked with this project to create confidence in the new servers' performance. I used qsql to calculate the average, max, min, and median speeds to compare with current standards. This information helped the company decide whether to purchase the servers.

For First Derivatives, I've created a chat room in q/kdb+, just like AIM or facebook messenger. One could create users to be authenticated against a database of usernames and encrypted passwords. Once logged on, the user would interact with an API I created to send messages to particular users or broadcast messages to all users. A daily log was also kept of global user status information – such as when a given user logged on or off for the day (HDB), what users were currently logged on (RDB), and status info including the ip address and port numbers used.

Lastly, regarding soft skills and technical guidance, I have always made a point of mentoring others and operating on a basis of open communication. I have taught developers more junior than me about the concepts employed in the abovementioned projects (I/O operations for the speed testing, the .z namespace for the chatroom, and general q basics like the functional/vector nature of q code, and breaking away from thinking in terms of for/while loops to name a few) and am happy to help whenever others are stuck or have questions. I have worked as a calculus tutor throughout college and worked in law enforcement for over a year, so working harmoniously as part of a team and providing mentorship wherever needed has always been important to me.

Experience

Work Experience

First Derivatives | New York, NY | Q/KDB+ Developer| 40 hours per week 10/2021 – Present

Created a database using q/kdb+ to store the disk read/write speeds for several Linux computers. QSQL was performed on this database to calculate statistical metrics like average, max, min, and median speeds, and compared with a benchmark. This informed the business if it is worthwhile to keep these Linux servers or return them to the original vendor.
Created a chat room in q/kdb+ holding a database of usernames/passwords that authenticates users when signing in, using inter-process communication (IPC) that leverages TCP/IP.
Developed and maintained a large q/kdb+ codebase. Supported several kdb apps. Used Linux to manage q processes.
Created a bash script in Linux to produce a directory tree from an input textfile describing the directory.
Completed a sample Git project covering the basics of pushes, pulls, merges, and branching.
Performed QA testing for the Canadian Security Authority Market Analysis Platform (CSA MAP), which monitors and investigates suspicious trading activity.

Roosevelt Island Operating Cooperation | New York, NY | NYS Special Patrolman | 40 hours per week 10/2018 – 12/2019

Responded to emergency calls to protect persons or property from crimes such as trespassing, larceny, and assault.
Regulated traffic during emergency FDNY/NYPD responses to hazards, criminal activity or traffic accidents.
Wrote incident reports for the abovementioned anomalies.

Baruch College | New York, NY | Calculus Tutor, Teaching Assistant | 5 hours per week 9/2013 – 12/2015

Taught algebra, trigonometry, differential calculus, integral calculus, and C++ programming to undergraduate students at Baruch College. Met with students in groups of 1 to 3 for 1 to 2 hours.
Acted as teaching assistant for a college algebra course. Substituted for the adjunct professor whenever he was absent. Helped facilitate classroom activities like grading papers and coordinating the distribution of class materials.

Internships

Genworth Financial | Richmond, VA | Actuarial Intern| 40 hours per week 5/2015 – 8/2015

Translated and rewrote a reserve calculation algorithm from APL to Excel's VBA. A vector programming language, APL, was used to calculate how much money should be kept in reserve to have confidence in the company's ability to pay out to it's life insurance policy holders. However, APL is a language that does not make use of keywords, and requires a separated keyboard of mathematical symbols to code in. This is problematic for most actuaries who cannot code in APL but depend on the output of this program. I was tasked with learning APL, decoding the program, and translating it to Visual Basic, so that the code may be read, understood, and edited within Microsoft Excel, which most actuaries do understand. This project was presented to 100+ employees including senior actuaries.

Wrote VBA code to create and maintain an inventory of over 400 projection reports and automate data extraction. Projections reports are critical to actuarial work. There was manual Exc⌈el work involved in producing each individual report and obtaining a summary of the whole collection. I used VBA to begin automating this process.

CitiMortgage | New York, NY | Risk Intern | 40 hours per week 7/2012 – 8/2012

Used VBA (Visual Basic) to write a program that automated the production of a matrix of pivot tables which described the relationship between the LTV (Loan-to-Value) of a person’s mortgage and that person’s FICO score in order to determine which loans were most likely to default.

Skills / Achievements

Proficient in R,Q,Python (PCAP certified), SQL, Unix, Git
Holds AWS Solutions Architect Associate cert
Currently working toward the Java OCA cert
Passed actuarial exams P and FM

Volunteer activities

LGR (LET’S GET READY) – New York, New York (6hrs/week) 1/2014 – 5/2014

Taught SAT math to a group of 10 low income high school students weekly, raising their scores on the SAT math section on average by over 100 points from the start to the end of the program.