Open to senior & applied-AI roles · Bengaluru
I build AI systems that ship.
Moneshwar Chandrasekar — Software Engineer in R&D at Testsigma. I turn uncertain AI workflows into fast, observable, production-grade systems across applied AI, agent infrastructure, backend, and full-stack.
01 · Selected work
Production work at Testsigma.
Three examples of turning ambitious product ideas into systems that real teams use every day.
01AI architecture · Testsigma
Taking the LLM out of the hot path
Replaced two sequential LLM passes with a deterministic mapper over recorded user actions. The system trusts what actually happened, and only falls back to a bounded, parallel LLM tier for actions it cannot observe.
7 secavg. generation, down from 4–5 minutes
- Python
- LLM systems
- Event streams
- Android adapter
02Agent infrastructure · Testsigma
Replacing polling with push
Re-platformed the browser and mobile runtime onto a new Python agent — WebSocket control and events, H.264 device video, manifest upgrades, operator-guided resume, iOS and cloud-device communication.
30 FPSH.264 streaming, up from 1 frame / 2–3 s
- Python
- WebSockets
- H.264
- scrcpy
- Appium
03Enterprise platform · Testsigma
Two products, one experience
Drove a large cross-product integration across backend, identity, and two React frontends — project linking, event-driven test creation, folder mirroring, silent SSO, and single-invite provisioning.
5+ systemsjoined into one product experience
- Java
- React
- Identity
- Event-driven
02 · Experience
Where the numbers come from.
3+ years across AI services, agent runtimes, and enterprise SaaS at Testsigma, Bengaluru.
Software Engineer · R&D
@ Testsigma · Bengaluru- Re-architected AI test-step generation from two sequential LLM passes to a deterministic mapper over recorded actions — down from 4–5 minutes to ~7 seconds, with zero LLM cost on the best-case path and a bounded model fallback for unobservable actions.
- Re-platformed the browser and mobile automation runtime onto a new Python agent: WebSocket control and events, H.264 video at 30 FPS (vs. one screenshot every 2–3 seconds), manifest-based upgrades, iOS and cloud-device support.
- Built an agentic test-authoring service from its first commit — exploration sessions, an LLM gateway, artifact storage and publication, bulk updates, and acceptance of learned test cases into the product.
- Delivered AI failure analysis and self-healing used across the customer base at 30K requests/day — failure classification, causal prior-step detection, ranked suggestions, and self-healing decisions powered by RAG, Qdrant, and tool calling.
- Drove a cross-product platform integration across backend, identity, and two React frontends: project linking, event-driven test creation, folder mirroring, silent single sign-on, and one-invite multi-product provisioning.
- Cut enterprise test-authoring load from 10s to 2s with caching, Hazelcast, JPA specifications, and query refactors; shipped Server-Sent Events across FastAPI, Java, and React for 4× faster AI feedback; remediated critical XSS findings.
- Java
- Python
- LLM systems
- Agent runtimes
- React
- AWS
Developer Intern · Full Stack
@ Testsigma · Bengaluru- Built core product surfaces from the ground up — the end-to-end Dashboard and 10+ Settings & Billing experiences spanning React and Java/JPA services; resolved P0 customer issues.
- Expanded natural-language test automation across Android, iOS, and web, and introduced controlled parallel execution to stabilize local test runs.
- Java
- Spring Boot
- React
- MySQL
03 · Toolkit
Built across the whole stack.
Comfortable moving from Java and Python services through realtime protocols, AI systems, and React surfaces.
- Languages
- JavaPythonTypeScriptJavaScriptSQLC++
- AI systems
- RAGLangChainQdrantTool callingAgent memoryLLM gateways
- Backend
- Spring BootFastAPINestJSJPAGraphQLRESTSSEWebSockets
- Frontend
- ReactReact QueryBrowser extensionsRealtime UIs
- Data & platform
- MySQLPostgreSQLMongoDBRedisHazelcastAWSS3DockerCI/CD
- Automation
- AppiumSeleniumscrcpy / H.264Test design
04 · Beyond the code
Build. Explain. Enable.
Speaking, awards, and leadership outside the day-to-day build.
Distinguished Speaker · Developers Conclave
Spoke to AI & ML students at Dayananda Sagar College of Engineering on real-world software engineering, AI advancements, and what the industry expects from young builders.
LinkedIn post ↗Agentic Summit, Bengaluru
Represented Testsigma, demonstrating its AI testing agent and seven specialized assistants to founders, builders, and AI teams.
Runner-up · AI in Test Automation
Led Team Himawari to build an AI Performance Testing Engine in a two-day sprint — judged by the co-founder of Portkey.ai and Testsigma's CEO.
Ownership Award, Q3
Recognized for exceeding delivery commitments with zero known bugs.
Sole technical lead · vibe-coding contest
Independently supported ~100 non-technical participants through Testsigma's internal vibe-coding contest — troubleshooting and unblocking throughout.
05 · Independent work
Curiosity, shipped.
Personal systems built to scratch real itches and test new stacks.
01News Intelligence RAG PlatformThree-service pipeline for news ingestion, semantic retrieval, and streaming conversational search.Python · Express · Qdrant · Redis · React↗02Multi-Tenant SMS GatewayAuthenticated device orchestration with SMS delivery-state tracking through physical SIM cards.FastAPI · PostgreSQL · React↗03TNEA Data LensFast, filterable exploration of Tamil Nadu engineering admissions data.Next.js · TypeScript↗
Wide surface area.
One engineering mindset.
I’m a Software Engineer in R&D at Testsigma with 3+ years of experience across AI-assisted test authoring, browser and mobile agent runtimes, and enterprise SaaS. I work end-to-end — from Java and Python services through realtime protocols, infrastructure, identity, and React product surfaces.
I author architecture RFCs, drive cross-repository releases, and help engineers build a clear mental model of the system — onboarding new hires and mentoring interns along the way.
“Use AI where uncertainty is real. Use software where the facts are already known.”
- Education
- B.E., Electronics & Communication — Sri Venkateswara College of Engineering, Chennai
- Years
- 2020 – 2024 · CGPA 9.19 / 10
Contact
Let’s build something that holds up.
Open to senior software, applied AI, backend, full-stack, and agent-infrastructure roles.