A Week of Testing AI Career Tools While Learning the Sitar

On Monday evening, when I placed the sitar on my lap after work, the tip of my left index finger was throbbing. I had spent the entire day typing away at the keyboard, comparing and evaluating three AI career development tools. When I pressed a string, the wrong note rang out, and my teacher said from the other side of the screen, "Your fingers need to be listening on the strings." That one remark strangely overlapped with what my team lead had said in the meeting room that afternoon: "Don't just read the results the tool spits out. Listen for the patterns."

This week, the company had scheduled an evaluation of five candidate AI tools for internal employee career path recommendations. As a security analyst, my share of the work covered data privacy and access permissions, but I ended up also reviewing actual usability alongside the HR team. At the same time, the online sitar lessons I had started two months ago were scheduled for Tuesday and Thursday evenings. October in Bengaluru still brings occasional rain, and if you leave the windows open, the sound of practice leaks out between the raindrops.

Tuesday: Practising How to Listen for Patterns

I spent the entire morning testing the second tool. This one worked by taking an employee's past project history and tech stack as input and suggesting three suitable roles for the next 18 months. It had a screen that visually displayed the reasoning behind each suggestion, letting you see which skills connected to which roles via a graph. I entered my own history as a test. The suggestion came back: "Given your experience in threat modelling, consider a cloud architecture security lead role." It wasn't wrong, but it completely failed to reflect why I find fulfilment in my current role. It read the numbers and keywords well but missed the texture of context.

During the evening lesson, my teacher slowly walked me through the first four notes of Raga Yaman. "Play one note and listen until it fades away," he said. "The next note comes only after that." I kept trying to reach for the next note too early and threw off the rhythm. It was the same when evaluating the AI tools. Instead of rushing to judge the moment results appeared on screen, I should have been tracing the path back to what data those results came from. I was too impatient.

Wednesday: The Gap Between Expectations and Reality

The third tool had a feature where you could type in natural language, "What kind of work do I want to be doing in three years?" and it would analyse your current skill gaps and generate a learning roadmap. It was impressive in the demo video, but when it came to recommending certifications or training courses relevant to the Indian IT market, its limitations showed. The certification lists were mostly U.S.-centric, and there were hardly any options available for in-person attendance in Bengaluru. From a security standpoint, the career aspirations employees enter are themselves sensitive data, yet this tool's data retention policy amounted to a single page of documentation, so I drew a red line in my notes.

At lunch, I ate bisibele bath at a small restaurant near Indiranagar and talked with my colleague Ananya. Ananya is on the HR team and was sceptical about whether employees would actually trust tools like these. "People want to have a conversation with their manager. They don't want an algorithm telling them, 'You should go this way.'" She was right. A tool shouldn't replace conversation; it should serve as a starting point for one.

Thursday: What the Fingers Remember

The fourth tool was the type that integrates with an existing LMS (Learning Management System). It was the most technically stable, and its API documentation was clean. However, the UI was overly simplistic, too flat a screen to handle something as complex as career pathing. It passed most items on the security checklist: encryption methods, granular access permissions, presence of audit logs. Using this tool would make things easy for our team, but the concern was that the employee experience might feel bland.

Thursday evening lesson. This time I learned meend, the technique of sliding your finger along the string to bend a note. At first I couldn't control the pressure, so the note would shoot too high or cut off midway. My teacher said, "This isn't a guitar. On the sitar, the amount of time your finger stays on the string determines the sound." Security analysis works the same way. When scanning logs, if you spot a single anomaly, you shouldn't just pass over it. You need to linger at that point for the pattern to emerge. Rush to the next log entry and you'll miss it.

Around 10 p.m., I finished practising and stepped out onto the balcony. Lights glittered in the direction of Koramangala, and somewhere an auto-rickshaw horn blared long and loud. The marks from the strings were still clearly etched into my fingertips. Tomorrow I would need to test the fifth and final tool.

Friday: Writing the Report

The fifth tool had the most polished interface, but at over ₹800 per employee per month, the price was nearly double that of the other tools. In terms of features, it was the most comprehensive, including skill mapping and mentor matching. From a security standpoint, though, I noticed a known vulnerability pattern in its SSO integration and sent an inquiry email to the vendor. I needed their response before I could finalise the evaluation.

In the afternoon, the whole team gathered for a mid-point review. I summarised my security assessment and created a comparison table showing how each tool handles employee data. The tool with the strongest security didn't necessarily deliver the best user experience. Conversely, the tool with the sleekest UI had an opaque data retention policy. There was no perfect option, only a question of which values to prioritise.

On Friday evening, there was no lesson, but I practised the aroha (ascending scale) of Raga Yaman on my own for 30 minutes. Ni Sa Re Ga Ma Pa Dha Ni Sa. After playing the same notes about fifteen times, my fingers began to remember the positions, and there were brief moments when my hands moved before my mind could think. The week of repeatedly tinkering with five AI tools had been similar. At first, each tool's interface felt unfamiliar, but by Friday, the moment I logged in, my hands already knew where to click.

Looking back on the week, I'm not trying to force a connection between AI career tools and the sitar. It's just that both demanded the same thing of me: whether it's sound or data, the willingness to dwell within the results rather than rushing to judge them. There were things that could only be heard when fingertips lingered long enough, on the strings or on the keyboard. Next week, I'll receive the vendor's response and finalise the report, and on Saturday morning, I plan to start the new raga my sitar teacher introduced.