Official Resume • Rishav Kumar (Cognizant FTE)
Portfolio Website

Rishav Kumar

Software Development & Automation Engineer • Programmer Analyst at Cognizant

Kolkata, West Bengal, India +91 7903017087 rishavkumara13@gmail.com rishav-kumar.pages.dev linkedin.com/in/13rishav github.com/itsrishav13

Professional Summary

Results-driven Software Development & Automation Engineer (Programmer Analyst at Cognizant) and technical product owner specializing in high-throughput data validation engines, autonomous CI/CD pipelines, and enterprise automation utilities. Primary Product Owner & Lead Developer of iMigrator across 60+ enterprise projects, re-engineering core comparison pipelines with Polars LazyFrames and DuckDB to achieve 3x throughput and zero-OOM execution on 50M+ row tables. Proven leadership in integrating GenAI (AWS Bedrock / Anthropic Claude) for automated root-cause diagnostics, designing atomic hot-swap self-updaters with PyInstaller, and orchestrating collision-safe Jenkins pipelines across 17+ database platforms and 12+ data formats.

Professional Experience

Cognizant Technology Solutions 03/2025 – Present
Programmer Analyst — Lead Automation Engineer & Product Owner 08/2026 – Present (Promoted: 7 Aug 2026)
  • Platform Ownership & Scale: Primary Product Owner and Lead Developer of iMigrator (v6.4 → v6.6/v7), scaling validation and automated reconciliation across 60+ enterprise client projects using generic, reusable architecture.
  • High-Performance Engine Migration: Re-architected core comparison from Pandas memory-bound execution to Polars LazyFrames and DuckDB in-memory relational engine, delivering 3x faster processing and a 60%+ memory reduction.
  • Adaptive Memory & Chunk Streaming: Implemented psutil dynamic RAM allocation (70–85% RAM) and disk-backed chunk streaming (fetch_to_disk), enabling low-memory systems (8GB/16GB RAM) to process 50M+ records across 50 columns in 10–15 minutes with zero OOM errors.
  • Cross-Database Join & Snowflake Pushdown Engine: Engineered reconciliation engine staging heterogeneous data into temporary Snowflake tables, executing distributed joins directly in-warehouse to eliminate multi-terabyte network data transfer.
  • GenAI Diagnostics & Multi-DB Connectors: Architected autonomous AWS Bedrock & Claude failure diagnostics with pattern clustering and structured JSON remediation; built query adapters across 17+ platforms (Snowflake, Fabric, DynamoDB, MongoDB, DB2) and 12+ formats.
  • Enterprise Packaging & Jenkins CI/CD: Designed 3-tier zero-downtime hot-swap updater via PyInstaller (Launcher → Updater → Real App) with UNC share sync; deployed collision-free parameterized Jenkins pipelines with Lockable Resources and Rally automation.
Programmer Analyst Trainee — ETL Testing & Automation Developer 08/2025 – 08/2026
  • Top 5-Star Appraisal Rating & Promotion: Conferred top 5-Star (“5 - High”) performance rating across all 13 core evaluation competencies in 1st-Year Confirmation Appraisal; commended by engineering leadership for rapid Python mastery and pivotal contributions to iMigrator, accelerating promotion to Programmer Analyst on August 7, 2026.
  • Automation Team Transition (Subsequent Months): Following initial 7-month ETL testing, transitioned to the Automation Team as an Automation Developer, authoring modular Python test utilities, automated test case generators, and core reconciliation logic for iMigrator.
  • Enterprise ETL Testing & Verification (First 7 Months): Executed comprehensive source-to-target test verification, data quality audits, and migration testing; formulated complex SQL queries across Snowflake and Oracle targets, validating transformations and constraints; logged and tracked defects in CA Agile Central (Rally).
Data Engineering Intern — Informatica & ETL Foundations 03/2025 – 07/2025
  • Healthcare Data Integration Pipeline: Built end-to-end Informatica PowerCenter ETL pipeline ingesting multi-source insurance feeds (Group, Subgroup, Subscriber flat files) into dimensional Oracle tables using Expression, Filter, Router, and Joiner transformations.
  • XML Publishing & Workflow Management: Developed automated downstream publishing workflow generating XSD-compliant personalized XML letters; configured and scheduled session runs in Informatica Workflow Manager.
Page 1 of 2 • Scroll below for Page 2 ↓
Rishav Kumar • Professional Resume Page 2 of 2

Featured Projects

iMigrator Platform Evolution (v6.4 → v6.6/v7)
Python 3.12 • Polars • DuckDB • Snowflake Pushdown • Multi-DB Connectors (17+ Platforms) • AWS Bedrock • Claude • Jenkins
  • High-throughput data validation platform adopted across 60+ enterprise projects; expanded core engine from 1,000 to 2,560+ LOC.
  • Engineered Snowflake pushdown compute modules generating temporary views to execute reconciliation directly in-warehouse, eliminating multi-terabyte network data transfer.
  • Built automated MD5 hash keyless fallback reconciliation, extraction adapters for MongoDB/DynamoDB document schemas, and automated DML execution blockers to prevent accidental destructive SQL operations.
DataCraft — Autonomous SQL & Data Engineering Assistant
Python • Google Agent Development Kit (ADK) • Gemini API • Firestore • Snowflake • BigQuery • PostgreSQL • Vertex AI
  • Architected an autonomous multi-modal AI agent using the Google Agent Development Kit (ADK) during Google's 4-Hour "Build with Gemini" hackathon, earning the official Google Developer Badge & Credly Certification.
  • Engineered a dynamic Firestore schema catalog (list_tables, get_table_details, add_table) registering database schemas, primary keys, and data types across multiple database platforms.
  • Built automated SQL anti-pattern detection using sqlparse (flagging full scans, missing filters, and uncapped sorting) with dialect-specific optimizations for Snowflake, BigQuery, and PostgreSQL.
  • Implemented multi-modal generation tools producing visual ER diagrams via gemini-3.1-flash-lite-image and animated Kafka event-streaming architecture videos via gemini-omni-flash-preview on Vertex AI with GCS storage.
  • Integrated Vertex AI Memory Bank (PreloadMemoryTool) for session-level dialect persistence, paired with AgentEngineSandboxCodeExecutor and a responsive FastAPI/A2UI card interface.
Enterprise Packaging & Multi-Tier Self-Updater
Python • PyInstaller (onedir/onefile) • Windows API • SQLite • UNC Network Shares
  • Architected 3-tier zero-downtime hot-swap updater (Launcher → Updater → Real App) with build repair, rollback protection, and CLI maintenance utilities (IMIG_Updater.cmd).
Healthcare Payer Data Integration Pipeline
Informatica PowerCenter • Oracle SQL • XML / XSD • Flat File Processing
  • Built multi-source ETL pipeline processing fixed-width and delimited patient insurance data, cleansing and routing errors to exception tables, and generating personalized XML welcome letters.