General Information
- Roles: Senior Software Engineer
- Email: sanketkamath1996@gmail.com
- Phone: 9449063585
Skill | tools |
---|---|
Programming language | Python (primary), Rust, Javascript, Bash, C# |
Platforms | Linux, Windows |
Libraries used | Pandas, Pydantic, Sqlalchmey, Request, Fastapi |
Database | sqlite, postgres |
Editors | Vim, Vscode, Pycharm |
Orchestration | Kubernetes, Airflow, Azure pipeline, docker |
Feilds worked | Task Mining, Automation |
Others | pdb, ssh, git |
Company | Duration |
---|---|
SOROCO | 6.5 Years |
Work Projects
Business Mapping and Contextualization of Gaps
Utilized AI-powered tools to generate potential Key Performance Indicators (KPIs) and business problems related to industrial processes.
Contributions:
- Generate prompts to generate CSV formatted data for give industry and process.
- Handle token limit for the same and other issues for prompt output.
- Make notebook self-serviceable to reduce developer time.
Technologies Leveraged:
- OpenAI library to communicate with llama models.
- Pandas to parse the output.
- Postgres to store and retrieve data.
Digital Gap framework
Improved user experience by identifying gaps and suggesting solution for optimization of workflow.
Contributions:
- Design Frameworks as two-pass calculation approach:
- Python logic to generate evidence with task metrics.
- Utilized a configurable rule engine to evaluate task metrics meets conditions or mark gaps in the system.
- Implementation of rule engine and few evidences.
- Later Assist others to adopt framework for other type of gaps.
Technologies Leveraged:
- Pydantic to validate and parse rules.
- Airflow to scheduling workflow.
- Postgres to store and retrieve data.
Document Engine
Generate high level integration of click and keystroke events on excel to into meaningful actions like copy or format table.
Contributions:
- Design a two-step rule engine for grouping keystroke/click to meaningful action.
- first mark particular key stroke as components of action
- generate long string and match regex to find pattern of components which actually make action.
- Implement the rule engine.
- Provide rule like copy, format table etc.
- Later Assist others to adopt rule engine to evaluate action on other apps.
- It became core part of team responsible for action detection.
Technologies Leveraged:
- Pydantic to validate and parse rules.
- Pandas
pd.eval
to run complex operation of large dataset. - Regex for combining
- Airflow to scheduling workflow.
- Postgres to store and retrieve data.
Kubernetes Integration for platform for automation and task mining
Successfully integrated automation and task mining platform with Kubernetes and Cloud Platforms.
Contributions:
- Dockerization of existing service (phase-1)
- Design and write platform agnostics libraries various backend.
- Write Helm charts for Kubernetes based installation.
Technologies Leveraged:
- Kubernetes, helm for orchestration and installation.
- Terraform for provisioning.
- Various services for backend:
Service | AWS | Azure | Bare Metal |
---|---|---|---|
Secret Manager | AWS Secrets Manager | Azure Key vault | HashiCorp Vault |
Storage | S3 | Blob Store | Minio |
Database | RDS | Azure Postgres | Postgres |
Logging | ELK | ELK | File system |
Docs server setup
Create a service for documentation to be hosted on each deployment. Also deploy same documentation for marketing on existing word press server.
Contributions:
- Convert static documentation into php files to allow authenticated user of WordPress server to access it.
- Create service which serves static documentation behind ngnix in each deployment.
- Automate the process of updating and deploying services to allow self-service by the documentation team.
- Clean up duplicated static files reducing size by 50%.
Technologies Leveraged:
- Azure pipeline: To automate build of new service and deploying files to php server.
- WordPress: Integrating with documentation existing marketing pages.
Supply Chain Automation
Automated part of supply chain process for a pharmaceutical company.
Contributions:
- Design a framework to parse information from documents using two components:
- Each parser had to function:
- identify if it can parse given document.
- parse the given document.
- Orchestrator which registered and orchestrated parsing of formats.
- Each parser had to function:
- Implement parser to parse multiple formats of data.
Technologies Leveraged:
- Tessract OCR to extract data from searchable pdf.
- PyMuPdf to extra data from searchable pdf.
- Use pandas to extract data from excel.
- Django to server data for validation of extracted data.
Insurance Process Automation
Automated two business processes for an insurance company.
Contributions:
- Refactored existing code to use command pattern, increasing efficiency and reduced testing time.
- Automate complex web based process using selenium wrappers.
- Automate wc3270 terminal interaction by parsing screen.
Automation Framework Development
Established coding guidelines for automation projects.
Contributions:
- A template for automation system setup, utilizing static checkers to ensure code quality
- A dashboard for error analysis and retry automation processes from checkpoint.
- Wrapper libraries for handling secrets, databases, and logging
Business Transaction Reconciliation(Internship)
Reconciled complex business transactions for an e-commerce company over 6 months at Soroco.
Contributions:
- Reconciled transactions across multiple files, involving intricate business processes
- Successfully utilized Pandas to streamline reconciliation process
- Use pandas internal function for parsing html table which reduce deadline by month.
Technologies Leveraged:
- pandas for faster computation.
Education
Degree | Institution | Year | Grade |
---|---|---|---|
B.E in C.S. | RVCE | 2018 | 9.10 |
P.U. | Chetan P.U. | 2014 | 94.83 |
SSLC | Chetan School | 2012 | 99.12 |
Personal Project
Improving lsp support for vim-ale
Todo.conky (src)
Todo Widget for Linux desktop widget written in Lua to view todo list.