Vishal Krishna

AI Researcher

Redmond, Washington, United States12 yrs 6 mos experience

Most Likely To SwitchAI Enabled

Key Highlights

Expert in Security and Machine Learning solutions.
Led AI projects at Microsoft with significant impact.
Proven track record in optimizing LLM performance.

Stackforce AI infers this person is a Software Engineer specializing in Machine Learning and Security within the tech industry.

Contact

viscrisn@yahoo.co.in LinkedIn

Skills

Core Skills

SecurityMachine LearningCloudSoftware Engineering

Other Skills

AI PlatformAlgorithmsCC++Cloud ML migrationCloud transitionComputer ScienceCross-validationData AnalysisData StructuresDeep learningEclipseGraph algorithmsJavaLLM optimization

About

Experienced Software Engineer with a demonstrated history of working in the computer software industry. Skilled in Security, Machine Learning, Cloud, Backend engineering etc. Works on cutting edge generative AI solutions, carving the path for others to follow.

Experience

12 yrs 6 mos

Total Experience

3 yrs 1 mo

Average Tenure

3 yrs 5 mos

Current Experience

Microsoft

3 roles

Artificial Intelligence Engineer

Promoted

Jan 2023 – Present · 3 yrs 5 mos · New York, United States

Leading projects for Microsoft Security Copilot.
AI Platform and orchestrator for connecting LLM model and APIs together.
Responsible AI
Latency and cost optimization for LLMs
Novel techniques for LLM planner and response engine

SecurityAI PlatformLLM optimizationResponsible AIMachine Learning

Software Engineer

Jun 2017 – Jun 2020 · 3 yrs

Worked on Windows operating system handling security and performance. Billions of devices which runs on windows are updated each month to keep them secure and efficient.

Windows OSSecurityPerformanceSoftware Engineering

Software Engineer

Jul 2014 – Jun 2017 · 2 yrs 11 mos

Developed end to end security platform for entire Microsoft software quality check and distribution platform
Worked on windows 10 mobility platform to provide seamless developer experience across multiple devices.
Worked on creating middle tier services in mobility area for smooth transition from on premise to cloud by providing utilities, documentation, components and recommendations.
Created easy to use play and components for features such as authentication, instrumentation, encryption etc. to be used by simple package inclusion by developers.

Security platformWindows 10 mobilityCloud transitionSoftware EngineeringSecurity

Twitter

Machine Learning Engineer

Jun 2020 – Jan 2023 · 2 yrs 7 mos · Seattle, Washington, United States

Relevance/Ranking recommender system, Infrastructure, Cloud ML migration and ML tools.

Recommender systemsCloud ML migrationML toolsMachine LearningCloud

Indian institute of science

TagMe! - Tag images into one of five different categories

Mar 2014 – Mar 2014 · 0 mo · Bangalore

Used deep convolution network and few other algorithms to generate set of features, applied Random forests and SVM respectively creating hierarchical classier, fine tuned custom ensemble model
Secured rank 9 (out of 644) with algorithm accuracy of 96.4%

Deep learningRandom forestsSVMMachine Learning

Indian institute of technology, kharagpur

Research Assistant

Jun 2013 – May 2014 · 11 mos

Research Assistant and Intern in School of Medical Science and Technology

Indian institute of science

Twitminer - Classfier to classify tweets

Mar 2013 – Mar 2013 · 0 mo · Bangalore

Twitminer - Classier to classify tweets from twitter.com into dierent categories March 2013
Removed redundant information from tweets using C++, used combination of "bag of words" and 8F to extract features, applied Naive-Bayes classier using python and tuned threshold using cross-validation
Secured rank 15 (out of 118) with algorithm accuracy of 92.08%

C++Naive-Bayes classifierCross-validationMachine Learning

Facebook: machine learning

Recommend missing links in a directed social graph

Jun 2012 – Jan 2013 · 7 mos

Graphs approach similar to Page-Rank Algorithm. Calculated probability of nodes having backlinks the highest priority. Sorted and recommended most probable nodes based on connections. Parallel
implementation using MPI improving execution time
Rank 132 (out of 424), algorithm accuracy of 0.69267 (highest accuracy being 0.72411)

Graph algorithmsMPIMachine Learning