Industry Experience
3+ Years
Built backend, automation, cloud, and enterprise software across internships and full-time roles.
Open to opportunities
Distinct Value
I build AI-native developer tools, backend systems, and product surfaces that operate in the real world.
Software engineer with production experience across enterprise workflows, API platforms, AI agents, and cloud-backed systems. I care about safe execution, clear interfaces, and architecture that teams can ship.
Current focus:
Open to opportunitiesOracle internship completed Jul 2026M.Tech (AI & Data Science) @ NIT Durgapur
Case Studies
4 Deep Dives
Developer tools, AI agents, career infrastructure, and operations-focused engineering systems.
Public Profiles
5 Platforms
GitHub, LinkedIn, LeetCode, HackerRank, and X with public proof of work.
Keshav Kumar Jha
Software Engineer | AI Agents | Backend Systems
- Oracle Aconex internship completed in July 2026.
- Experience across backend APIs, AI agents, React, Electron, and cloud systems.
- Strong focus on reliability, performance, and practical user impact.
About
From enterprise systems to AI agents, I build for real constraints.
My edge is cross-domain thinking: I can reason about product UX, backend integration, automation safety, and system behavior together. That helps me build software that is useful, explainable, and reliable.
Experience Snapshot
Jan 2026 - Jul 2026
Project Intern @ Oracle Financial Services Software
Contributed to Oracle Aconex document-management workflows across enterprise frontend, backend, and cloud systems.
- Worked across Angular, Spring Boot, Oracle SQL, and Oracle Cloud Infrastructure.
- Built Playwright automation for document creation, metadata validation, hierarchy checks, uploads, and version management.
- Investigated production defects and improved release confidence through debugging, reviews, and regression coverage.
Aug 2022 - Oct 2025
Backend Developer and DevOps Engineer @ Distronix Pvt. Ltd.
Built and maintained backend services, operational dashboards, monitoring systems, and Linux-hosted deployments.
- Designed Node.js and Express APIs for employee management, payroll, inventory, procurement, and monitoring workflows.
- Containerized services with Docker and deployed production applications across Linux servers.
- Worked across PostgreSQL, MySQL, MongoDB, dashboard integrations, logging, and production issue resolution.
Education Context
2024 - 2026
M.Tech in AI & Data Science
NIT Durgapur
Focused on machine learning, data systems, and practical deployment patterns for real-world applications.
2018 - 2022
B.Tech in Computer Science
University of Engineering and Management, Kolkata
Built core foundations in software engineering, data structures, and system design.
How I Build
Engineering approach that shows how I think, not just what I ship.
Hiring teams care about decisions and tradeoffs. These are the principles I use when building product features, backend systems, and AI-assisted workflows.
Build Around Real Workflows
I start with the actual task a developer or operator needs to finish, then shape the system around it.
Tradeoff
It takes more discovery up front, but the final product solves the right problem.
Applied Example
EchoMind and The REST Project both prioritize tool execution, context, and approval paths over surface-level chat.
Keep Systems Explainable
When AI or automation is involved, I make actions, evidence, and uncertainty visible.
Tradeoff
Adds product and data-model complexity, but it creates trust and debuggability.
Applied Example
StackCendra is planned around evidence-backed facts, confidence scores, and human-approved operations.
Design For Safe Execution
I prefer narrow tools, permission boundaries, logs, and validation before automation touches real systems.
Tradeoff
Safer flows can feel slower at first, but they prevent costly hidden failure modes.
Applied Example
EchoMind uses a typed tool registry, risk-tiered policy engine, approvals, and audit logs before running actions.
Project Case Studies
Not just screenshots. Problem, decisions, and impact.
Each project answers what problem it solves, what I owned, and what technical decisions made it work.
The REST Project
A Postman-inspired developer workspace for APIs, databases, teams, SSH, and local terminal workflows.
Problem
API development often scatters requests, environments, database checks, SSH sessions, and team context across separate tools.
My Contribution
I built the web and Electron product surfaces, request workflows, database integrations, workspace model, and local developer capabilities.
Impact
Created a single workspace for REST and GraphQL testing, SQL exploration, shared collections, environments, authentication, and desktop-only terminal/SSH workflows.
Next.jsElectronTypeScriptFirebaseSQL DatabasesSSHDocker
EchoMind
A local-first desktop AI assistant with voice, memory, approvals, tool execution, and automation.
Problem
Most assistants can chat, but they do not safely operate a local machine, remember user workflows, or expose what they are doing.
My Contribution
I built the Electron desktop shell, FastAPI agent service, typed tool registry, permission system, voice pipeline, memory layer, workflows, and automation tools.
Impact
Completed the original eight-phase roadmap into a working personal computing platform with local STT/TTS, screen awareness, routines, recovery mode, developer workspaces, and smart-home integration.
PythonFastAPIElectronTypeScriptLangGraphSQLitePlaywright
CareerOS
An AI career workspace across web, desktop, browser extension, and mobile.
Problem
Job search and interview preparation are split across trackers, notes, resumes, coding tools, calendars, and browser tabs.
My Contribution
I designed the monorepo architecture and built the web app, desktop companion, browser extension, mobile app, shared types, Firebase flows, and AI-backed career workflows.
Impact
Delivered a multi-surface product covering application tracking, job capture, interview preparation, system design practice, coding judge workflows, learning, analytics, calendars, profiles, settings, and resume tooling.
Next.jsTypeScriptFirebaseElectronReact NativeExpoChrome Extension
StackCendra
An AI-native engineering workspace for project discovery, environment drift, and production recovery.
Problem
Engineering failures often live between code, configuration, local setup, deployments, infrastructure, and telemetry, while teams debug them through disconnected tools.
My Contribution
I am shaping the Phase 0 product contract, architecture, wiki, roadmap, visual prototype, and release plan for an evidence-backed local-to-production workflow.
Impact
Defined a focused wedge around intelligent project discovery and a phased platform roadmap that grows toward configuration intelligence, controlled deployments, incident diagnosis, and sanitized production-to-local reproduction.
Next.jsTypeScriptTauriRustGoPythonPostgreSQLOpenTelemetry
Skill Focus
Core strengths first, supporting tools second.
I prioritize depth in a few high-leverage areas instead of listing every technology I've touched.
Backend And System Design
Designing APIs, services, and data flows that stay maintainable as product scope grows.
Node.jsFastAPISpring BootREST APIsPostgreSQLDistributed Systems
AI-Native Developer Tools
Building practical AI workflows with memory, tool execution, approvals, and product-grade UX.
LLM APIsAI AgentsRAGLangGraphOllamaTool Calling
Product Interfaces
Shipping React, Next.js, Electron, and mobile surfaces for complex workflows without hiding the system.
ReactNext.jsTypeScriptElectronReact NativePlaywright
Supporting Tools
Delivery And Collaboration
Tooling I use to ship production-ready software in teams.
GitDockerKubernetesFirebaseVercelOCILinuxCI/CD
Trust Signals
Proof of work, not just claims.
If you want to validate output quickly, these links point directly to my activity, resume, and professional profile.
Quick Validation Links
Public Profiles
Explore coding activity, competitive programming, and professional background.
Contact
Let's build something useful, not just flashy.
Share your project context and I'll respond with clear next steps. I typically reply within 24 hours.